Research ArticleSalp Swarm Algorithm for Node Localization in WirelessSensor Networks

Huthaifa M Kanoosh1 Essam Halim Houssein 2 and Mazen M Selim1

1Faculty of Computers and Informatics Benha University Banha Egypt2Faculty of Computers and Information Minia University Minya Egypt

Correspondence should be addressed to Essam Halim Houssein essamhalimmuedueg

Received 27 July 2018 Revised 6 January 2019 Accepted 21 January 2019 Published 19 February 2019

Guest Editor Noradin Ghadimi

Copyright copy 2019 Huthaifa M Kanoosh et al +is is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

Nodes localization in a wireless sensor network (WSN) aims for calculating the coordinates of unknown nodes with the assist ofknown nodes +e performance of a WSN can be greatly affected by the localization accuracy In this paper a node localizationscheme is proposed based on a recent bioinspired algorithm called Salp Swarm Algorithm (SSA) +e proposed algorithm iscompared to well-known optimization algorithms namely particle swarm optimization (PSO) Butterfly optimization algorithm(BOA) firefly algorithm (FA) and grey wolf optimizer (GWO) under different WSN deployments +e simulation results showthat the proposed localization algorithm is better than the other algorithms in terms of mean localization error computing timeand the number of localized nodes

1 Introduction

Wireless sensor networks (WSNs) are networks that consistof a few autonomous sensor nodes that are distributed in aspecific area for cooperatively sensing their environment Inthe last decade WSNs have been employed in many real-lifeapplications such as healthcare home automation trafficsurveillance and environmental monitoring [1]

Accurate node localization is a critical issue in WSNsLocalization problem in WSNs means calculating the po-sitions of unknown sensor nodes In many environmentswhere the sensor nodes of a WSN are deployed people maynot be able to go and fix these nodes In these environmentsthe sensor nodes are usually randomly scattered in randomlocations therefore the sensor nodes usually take randompositions On the other hand in many applications theinformation collected by the sensor nodes of a WSN may beuseless if the positions of the sensor nodes that gathered theinformation are unknown +is emphasizes the need for anaccurate node localization scheme [2]

In order to localize the sensor nodes of aWSN a GPS canbe attached to each sensor during the network deployment

+en the GPS can be used to find the coordinates of thesensor nodes However using GPS for localizing sensornodes is undesirable and impractical solution because ofmany reasons such as cost inaccessibility sensor nodes maybe deployed indoors and climatic conditions may disturbthe GPS reception ability [2] An alternative approach is toattach GPS to some nodes which are called anchor nodes orbeacon nodes +us the positions of these nodes with at-tached GPS are known after deploying the nodes of thewireless sensor network Using the known locations ofanchor nodes localization algorithms can be employed toestimate the positions of the unknown nodes [2] +ere aretwo classes of non-GPS based localization algorithmsnamely rang-free and range-based algorithms [3] Angle-based or point-to-point distance estimation among thesensor nodes is used with rang-based localization algo-rithms In these algorithms the positions of sensor nodes arecalculated by the assist of anchorrsquos trilateration [3] Unlikerange-based localization algorithms range-free localizationalgorithms do not need range information to estimate thepositions of the unknown nodes It only depends on thetopological information Compared to range-free localization

HindawiJournal of Computer Networks and CommunicationsVolume 2019 Article ID 1028723 12 pageshttpsdoiorg10115520191028723

algorithms range-based localization algorithms can achievehigher localization accuracy but at the expense of cost [4]

Recently node localization in WSNs is handled as amultimodal and multidimensional optimization problemthat can be solved using population-based stochastic ap-proaches In the literature many metaheuristic algorithmshave been employed to solve the localization problem inWSNs +ese algorithms have succeeded in reducing thelocalization error dramatically +ese algorithms attempt tosolve an optimization problem by trial and error in whichthe feasible solutions are processed and the nearest optimalsolution is identified Currently various optimization al-gorithms such as genetic algorithm (GA) cuckoo search(CS) gravitational search algorithm (GSA) butterfly opti-mization algorithm (BOA) particle swarm optimization(PSO) artificial bee colony (ABC) etc have been employedeffectively in specifying the positions of the unknown nodesin WSNs [5]

+e main contribution of this paper is using the SalpSwarm Algorithm (SSA) for the first time ever to localize thenodes of WSNs +e performance of the proposed SSA-basedlocalization algorithm is analyzed and compared with particleswarm optimization (PSO) butterfly optimization algorithm(BOA) firefly algorithm (FA) and grey wolf optimizer(GWO) algorithms +e results have shown that the SSA-based localization algorithm is better than the previouslymentioned localization algorithms in terms of computingtime number of localized nodes and localization accuracy

+e remaining sections of the paper are organized asfollows Section 2 covers some of the research efforts whichhave been done in the field Section 3 presents a brief de-scription for the different swam algorithms employed in thiswork Section 4 introduces the proposed SSA-based local-ization algorithm Section 5 includes the conducted ex-periments and results analysis Finally the paper isconcluded in Section 6

2 Literature Review

In the last years many optimization algorithms have beenemployed for addressing the problem of node localization inWSNs [6] In this section some of the recent relevant worksare covered and briefly described

Low et al have introduced a low-cost localization systemthat depends on the measurements obtained from a pe-dometer and communication ranging among adjacentnodes +e proposed system does not require good networkconnectivity and presents good performance in sparsenetworks A probability-based algorithm that needs anonlinear optimization problem solving is employed toprovide the localization information Moreover the particleswarm optimization (PSO) has been employed to determinethe optimum location of the sensor nodes Experimentalresults have proved that the proposed system has a goodperformance [7]

Manjarres et al have presented a hybrid node locali-zation algorithm based on Harmony Search (HS) algorithmwith a local search procedure +e main objective of theproposed algorithm is to address the localization problem

and to distribute its burdens over an iterative processAdditionally the proposed algorithm employs certainconnectivity-based geometrical constraints to decrease thearea in which each node can be located +e simulationresults have verified that the proposed algorithm is betterthan another simulated annealing- (SA-) based localizationalgorithm in terms of both localization error and compu-tational cost [8]

Li et al have suggested a self-adaptive artificial beecolony (SAABC) node localization algorithm that considersthe whole effects resulting from employing dynamic to-pology +e proposed algorithm provides good performancein WSNs with both random distributing nodes and dynamictopology Additionally the obtained simulation results haveshown that the proposed localization algorithm providesbetter node localization precision and precision stabilitycompared to the DV-Hop algorithm without the need forextra devices or overhead in communication [9]

Tamizharasi et al have proposed a novel hybrid nodelocalization algorithm based on bacterial foraging algo-rithm (BFA) and particle swarm optimization (PSO) +emain design objectives of the proposed algorithm are toenhance the efficiency and accuracy of BFA and to avoidgetting stuck in a local extreme In the proposed algorithmPSO is merged into the chemotaxis of BFA to speed up theconvergence rate Moreover the global search ability isimproved by proposing the elimination probability inelimination-dispersion based on the energy of bacteria +eobtained simulation results have verified that the proposedhybrid algorithm outperforms the BFA [10]

Tang et al have proposed a sensor nodes localizationalgorithm which depends on a new intelligent optimiza-tion algorithm called plant growth simulation algorithm(PGSA) that simulates the growth of plants In their workthey proposed inserting the plant root of adaptive backlightfunction into the original PGSA for improving conver-gence time and localization precision +e obtained sim-ulation results have verified that the proposed algorithm isbetter than the simulated annealing algorithm (SAA) interms of localization accuracy and computing time [11]

Jegede and Ferens have employed the genetic algo-rithm (GA) for learning the environmental issues within aWSN for effectively localizing its sensor nodes For everycoordinate in the grid network area given random per-turbations of received signal strength (RSS) GA would beable to learn the environment and to decrease the possibleerrors associated with the RSSI measurement taken foreach coordinate +e conducted simulation shows that GAcan reasonably localize sensor nodes using the co-ordinates of three anchors [12]

Goyal and Patterh have proposed a cuckoo search-(CS-) based node localization algorithm for estimating thecoordinates of sensor nodes in WSNs In the proposedalgorithm no weight coefficient is employed for con-trolling the global search ability +e conducted simula-tion has shown that the proposed localization algorithm isbetter than the particle swarm optimization (PSO) andvarious variants of biogeography-based optimization(BBO) in terms of localization accuracy [13]

2 Journal of Computer Networks and Communications

Dan and Xian-bin have presented a distributed two-phase PSO algorithm for efficiently and precisely localizingthe sensor nodes in addition to solving the flip ambiguityproblem In the first phase the initial search space is definedusing the bounding box method In the second phase arefinement process is performed for correcting the errorresulting from the flip ambiguity Additionally the proposedalgorithm attempted localizing sensor nodes that have onlytwo references or three near-collinear references +econducted simulation has proved the effectiveness of theproposed algorithm [14]

Krishnaprabha and Gopakumar have proposed a nodelocalization algorithm based on gravitational search algo-rithm (GSA) In the proposed work node localization inWSNs is formulated as a nonlinear optimization problemAlso the proposed algorithm tried to handle the flip am-biguity problem and to localize the sensor nodes that col-linear anchor nodes through the refinement phase +eobtained simulation results have shown that the proposedlocalization algorithm has good performance [15]

Peng and Li have focused on range-free localization as acost-effective alternative compared to range-based ap-proaches However they noticed that range-free localizationsuffers from higher localization error compared to the range-based algorithms In order to deal with this problem theypresented an improved version for DV-Hop which is apopular rang-free approach that depends on hop-distanceestimation +e improvement in the DV-Hop algorithm isperformed based on a genetic algorithm Simulation resultshave shown that the proposed localization algorithm hasbetter localization accuracy compared to other localizationalgorithms [16]

Sai et al have presented a hybrid node localizationalgorithm in WSNs which depends on the measurementsof the received signal strength (RSS) and parallel fireflyalgorithm (PFA) Taking into consideration the distancefactor the proposed algorithm transforms the node lo-calization problem in WSN into a nonlinear un-constrained optimization problem that is defined by anenhanced objective function In the proposed algorithmPFA estimates the coordinates of sensor nodes using thedistances between a sensor and a few numbers of its 1-hopneighbors Simulation results have shown that the pro-posed approach is better than PSO GA PFA and RSS interms of localization accuracy [17]

Sivakumar and Venkatesan have provided two-phasenode localization algorithm in WSNs In the first phase thepositions of sensor nodes are roughly estimated using arange-free localization method called mobile anchor po-sitioning (MAP) In the second phase the proposed al-gorithm attempts to reduce the localization error byemploying a certain metaheuristic approach In their workto perform the second phase they employed bat optimi-zation algorithm (BOA) modified cuckoo search (MCS)and firefly optimization algorithm (FOA) resulting in threelocalization algorithms namely BOA-MAP MCS-MAPand FOA-MAP +e experimental results have shownthat FOA-MAP is better than both BOA-MAP and MCS-MAP in terms of root mean square error (RMSE) [18]

Sun et al have proposed a multiobjective node locali-zation algorithm based on multiobjective particle swarmoptimization called multiobjective particle swarm optimi-zation localization algorithm (MOPSOLA) +e multi-objective functions include the space distance and thegeometric topology constraints In the proposed algorithmthe size of the archive remains limited using a dynamicmethod Simulation results have shown that the proposedalgorithm has achieved significant enhancements in terms oflocalization accuracy and convergence rate [19]

Arsic et al have proposed a node localization algorithmbased on fireworks algorithm (FWA) +e proposed al-gorithm provides optimal results in case of no rangingerrors and provides good results in case of ranging errorsMoreover they proposed an enhanced fireworks algorithm(EFWA) which achieved better results compared to FWAAlso the proposed localization algorithm outperforms theexisting PSO-based localization algorithm [20]

Shieh et al have compared several well-known heuristicssuch as genetic algorithm (GA) and particle swarm opti-mization (PSO) to more recent methods such as grey wolfoptimizer (GWO) firefly algorithm (FA) and brain stormoptimization (BSO) algorithms in terms of sensor nodeslocalization accuracy Also they proposed an enhancementin the localization algorithms to increase the number oflocalized nodes +e improved algorithms have been com-pared to the original ones in terms of the number of localizednodes and execution time in different network deployments[21]

Cheng and Xia have proposed a cuckoo search (CS)algorithm-based node localization algorithm In theproposed method the step size has been modified toobtain a global optimal solution in a short time Also thecandidate solutionsrsquo fitness is used for constructing mu-tation probability to avoid local convergence +e per-formance of the proposed algorithm has been evaluatedusing different anchor density node density and com-munication range in terms of average localization errorand localization success ratio +e simulation results haveproved that the proposed algorithm is better than thestandard CS and PSO in terms of average localizationerror and convergence time [22]

Daely and Shin have proposed a node localizationalgorithm based on dragonfly algorithm (DA) optimiza-tion algorithm +e proposed localization algorithm wasdesigned to determine the positions of the nodes whichare randomly distributed in a specific area +e simulationresults proved that the proposed DA based algorithm isbetter than PSO in terms of localization accuracy [23]

Nguyen et al have employed the multiobjective fireflyalgorithm to estimate the coordinates of sensor nodes inWSNs +e used objective functions depend on two criterianamely the nodesrsquo distance constraint and geometric to-pology constraint +e simulation results have shown thesuperiority of the proposed algorithm compared to well-known localization methods in terms of localization accu-racy and convergence rate [24]

Arora and Singh have used the butterfly optimizationalgorithm to localize the sensor nodes in WSNs +e

Journal of Computer Networks and Communications 3

proposed localization algorithm has been validated usingdifferent numbers of nodes with distance measurementscorrupted through the Gaussian noise +e simulation re-sults have shown that the proposed localization algorithm isbetter than several well-known localization algorithms in-cluding particle swarm optimization (PSO) and firefly al-gorithm (FA) in terms of localization accuracy [25]

Ahmed et al have presented a node localization algo-rithm based on whale optimization algorithm (WOA) whosemain objective is to localize sensor nodes in WSNs accu-rately [6] Also Sujatha and Siddappa have proposed ahybrid localization algorithm based on dynamic weightparticle swarm (DWPSO) and differential evolution (DE)algorithms +e authors observed that decreasing the squareerror of estimated and measured distance can improve thelocalization accuracy +e obtained simulation resultsproved that the proposed localization algorithm is betterthan the linearization method (LM) in terms of localizationaccuracy and performance stability [26]

Kaur and Arora have compared the performance ofseveral bioinspired algorithms including flower pollinationalgorithm (FPA) firefly algorithm (FA) grey wolf optimi-zation (GWO) and particle swarm optimization (PSO) inlocalizing the sensor nodes ofWSNs+e performance of thedifferent algorithms has been evaluated in terms of severalperformance metrics including localization accuracycomputing time and several localized nodes +e simulationresults have shown the superiority of FPA compared to theother algorithms in terms of localization accuracy [27]

To the best of our knowledge the Salp Swarm Algorithm(SSA) algorithm was never used for the localization probleminWSNs so far +erefore the main objective of this paper isto employ the SSA algorithm for handling the localizationproblem in WSNs and is to evaluate its performance againstseveral well-known swarm intelligence algorithms+e basicideas behind the SSA and other swarm algorithms are givenin the next section

3 Swarm Intelligence Algorithms

Swarm intelligence (SI) is a relatively new interdisciplinaryfield of research which has gained huge popularity in thesedays Algorithms belonging to this domain draw inspirationfrom the collective intelligence emerging from the behaviorof a group of social insects (like bees termites and wasps) Ithas successfully been applied to several real-world optimi-zation problems [28ndash30] In this section we review some ofthese algorithms that are employed in this paper in order tolocalize the nodes of WSNs

31 Particle Swarm Optimization Particle swarm optimi-zation (PSO) is a swarm optimization algorithm proposed byEberhart and Kennedy in 1995 It is inspired by the collectivebehavior of bird flocking and fish schooling It employsseveral particles that simulate a swarm population movingaround in the search space to find the best solution Eachparticle provides a candidate solution for the problem and isusually represented as a point in D-dimensional space +e

position of each particle is represented as a vectorxi (xi1 middot xi2 middot middot xi D) Particles move in the search spaceto find optimal solutions and each particle has a velocityrepresented as a vector vi (vi1 middot vi2 middot middot vi D) Duringmovement the position of each particle is updated based onthe best position that has been achieved by the particle itself(pbest) and the best position that has been achieved by all theparticles so far (gbest) +e velocity and the position of eachparticle are updated as shown below [31]

xt+1id x

tid + v

t+1id (1)

vt+1id wlowast v

tid + c1 lowast r1 lowast pid minusx

tid1113872 1113873 + c2 lowast r2 lowast pgd minusx

tid1113872 1113873

(2)

where t refers to the iteration number w is inertia weightthat aims at determining the effect of previous velocities oncurrent velocity c1 and c2 are acceleration constants r1 andr2 are random variables whose values are normally dis-tributed in [0 1] and pid and pgd indicate the elements ofpbest and gbest in the dth dimension respectively +e ve-locity is limited by a predefined maximum velocity vmax

32 Butterfly Optimization Algorithm Butterfly optimiza-tion algorithm (BOA) is a swarm intelligence algorithm thatis recently proposed by Arora and Singh [32] It is inspiredfrom the food-foraging behavior of butterflies Biologicallybutterflies find the source of food using sense receptors +erole of these sense receptors also called chemoreceptors isto sense fragrancesmell [33]

In BOA butterflies are the search agents that performoptimization Each butterfly emits fragrance with some in-tensity +is fragrance is propagated and sensed by otherbutterflies in the region +e fragrance emitted by butterfly iscorrelated with the butterflyrsquos fitness +is means that thefragrance of a butterfly changes according to its current lo-cation [32] When a butterfly is able to sense fragrance fromany other butterfly that is larger than its fragrance it will movetoward the latter and this phase is called global search On theother side when a butterfly cannot sense fragrance from otherbutterflies that is larger than its local fragrance it will moverandomly and this phase is called local search [34] In BOAthe fragrance is defined as a function of the physical intensityof stimulus [34] as shown below

fi cIa (3)

where fi is the perceived magnitude of fragrance c is thesensory modality I is the stimulus intensity and a is thepower exponent dependent on modality which accountsvarying degree of absorption As mentioned before there aretwo phases in the BOA algorithm namely local search phaseand global search phase In global search phase the butterflymoves one step toward the best butterflysolution glowast whichcan be formulated as

xt+1i x

ti + r

2times glowast minusx

ti1113872 1113873 times fi (4)

where xti is the solution vector xi of ith butterfly in iteration

number t glowast is the best solution found in the current stage

4 Journal of Computer Networks and Communications

fi is the fragrance of ith butterfly and r is a random numberin [0 1] Local search phase is formulated as

xt+1i x

ti + r

2times x

tk minusx

tj1113872 1113873 times fi (5)

where xtj and xt

k are jth and kth butterflies chosen randomlyfrom the solution space Equation (5) is considered a localrandom walk if and only if xt

j and xtk belongs to the same

subswarm and r is a random number in [0 1] Search forfood and mating partner by butterflies can occur at bothlocal and global scale therefore a switch probability p isemployed in BOA to switch between common global searchand intensive local search

33 Firefly Algorithm Firefly algorithm (FA) is a nature-inspired algorithm proposed by Yang and He [35] It mimicsthe social behaviors and flashing patterns of fireflies FAdepends on the next three idealized rules [36]

(i) Fireflies are unisex whichmeans that a firefly can getattracted to any other firefly regardless their sex

(ii) +e attractiveness of fireflies is directly proportionalto their brightness +us for any two flashingfireflies the firefly with less brightness moves to-ward the one with higher brightness If there is no abrighter firefly than a particular firefly the lattermoves randomly

(iii) +e brightness of a firefly is calculated using anobjective function

A fireflyrsquos attractiveness is proportional to the light in-tensity visualized by other fireflies in the region thereforethe relationship between the attractiveness β and the dis-tance r can be formulated as

β β0eminuscr2

(6)

where β0 is the brightness at distance r 0 and c is the lightabsorption coefficient +e movement of a firefly i toward amore attractive (brighter) firefly j is represented as

xt+1i x

ti + β0e

minuscr2ij x

tj minus x

ti1113872 1113873 + αtϵ

ti (7)

where rij is the distance between the fireflies i and j that iscalculated using the Euclidean norm the second term is dueto the attraction the third term is a randomization with αt

being the randomization parameter and ϵti is a vector thatincludes random numbers

34 Grey Wolf Optimizer Grey wolf optimizer (GWO) is arecent swarm intelligence algorithm inspired by the greywolf community It is developed by Mirjalili et al in 2014Grey wolf is a very dangerous creature which belongs to theCanidae family Grey wolves usually live in packs that consistof 5 to 12 wolves Each group has social dominance hier-archy alpha beta and omega in order +e alphas are amale and female which represent the leaders of the group+e betas are the second level of management hierarchy+eomegas are the final level in the hierarchy [37]

In order to mathematically model the social hierarchy ofwolves in GWO the fittest solution is referred to as the alpha(α) Consequently the second and third best solutions arebeta (szlig) and delta (δ) respectively +e rest of candidatesolutions are omega (ω) [35]+emathematical model of theencircling behavior is represented as follows [35]

Drarr

| Crarr

Xrarr

p (t)minus Xrarr

(t)|

Xrarr

(t + 1) |Xrarr

p(t)minus Ararr

Drarr

|(8)

where t indicates the current iteration Crarr

2 rrarr

2Ararr

2 ararr

rrarr

1 minus ararr X

rarrm is the position vector of the wolf r1

and r2 are random vectors in [0 1] and linearly varies from 2to 1 C and A are coefficient vectors and Xp is the positionvector of the prey During the optimization ωwolves updatetheir positions around αβ and δ +erefore the ω wolvescan reposition with respect to α β and δ as shown below[35]

Drarrα C

rarr1X

rarrαminus X

rarr11138681113868111386811138681113868

11138681113868111386811138681113868Drarrβ C

rarr2X

rarrβminus X

rarr11138681113868111386811138681113868

11138681113868111386811138681113868Drarrδ

Crarr

3Xrarrδ minus X

rarr11138681113868111386811138681113868

11138681113868111386811138681113868

Xrarr1 X

rarrαminus A

rarr1(D

rarrα)X

rarr2 X

rarrβminus A

rarr2(D

rarrβ)X

rarr3

Xrarrδ minus A

rarr3(D

rarrδ)

Xrarr

(t + 1) Xrarr

1 + Xrarr

2 + Xrarr

3

3

(9)

With these equations a search agent updates its positionaccording to α β and δ in an n-dimensional search space Inaddition the final position would be in a random placewithin a circle which is defined by the positions of αβ and δin the search space [37]

35 Salp SwarmAlgorithm Salps are part of Salpidae familywith the limpid cylinder design body +ey look like jelly-fishes in texture andmovement+e shape of a Salp is shownin Figure 1(a) +e water is pushed Salps bodies to moveforward [38] Generally the biological research about Salpsis still in its early stages because their living environmentsare hardly accessible and it is very difficult to keep them inlaboratory environment Salps swarming attitude is the maininspiration to build Salp swarm algorithm [39] Salpscompose a swarm in profound oceans which is called Salpchain +is chain is illustrated in Figure 1(b) +is chain canhelp to achieve better locomotion during the foragingprocess [40]

Originally the Salps population is divided into twogroups to formulate the mathematical model for Salp chainshead and followers +e head position is at the beginning ofthe chain while the rest of the chain is referred to as thefollowers [41]

+e Salps location is determined likewise swarm-basedmethods by an n-dimensional search area via considering

Journal of Computer Networks and Communications 5

the number of variables inside the presented problemrepresented by n Accordingly a two-dimensional matrixdescribed as x will reservation the position of all Salps It issupposed too in search space there is ldquoFrdquo which is a foodsource as the target of the swarm +e following equation issuggested to upgrade the leader location [37]

x1j

Fj + c1 ubj minus lbj1113872 1113873c2 + lbj1113872 1113873 c3 ge 0

Fj minus c1 ubj minus lbj1113872 1113873c2 + lbj1113872 1113873 c3 lt 0

⎧⎪⎨

⎪⎩(10)

where x1j shows the position of the first Salp (leader) in

the jth dimension Fj is the position of the food source in thejth dimension ubj indicates the upper bound of jth di-mension lbj indicates the lower bound of jth dimension andc1 cthe2 and c3 are random numbers

Equation (10) shows that the leader only updates itsposition with respect to the food source +e coefficient c1 isthe most important parameter in SSA because it balancesexploration and exploitation defined as follows [37]

c1 2eminus(4lL)2

(11)

where l is the current iteration and L is the maximumnumber of iterations

+e parameters c2 and c3 are random numbers uni-formly generated in the interval of [0 1] In fact they dictateif the next position in jth dimension should be towardspositive infinity or negative infinity as well as the step size Toupdate the position of the followers the following equationsis utilized (Newtonrsquos law of motion) [37]

xij

12

at2

+ v0t (12)

where ige 2 xij shows the position of ith follower Salp in jth

dimension t is time v0 is the initial speed and a vfinalv0where v x minus x0t

Because the time in optimization is an iteration thediscrepancy between iterations is equal to 1 and consideringv0 0 this equation can be expressed as follows [37]

xij

12

xij + x

iminus1j1113872 1113873 (13)

where ige 2 and xij shows the position of ith follower Salp in

jth dimension Using equations (1) and (4) the Salp chainscan be simulated Figure 2 shows the pseudocode to im-plement the SSA algorithm

4 Formulation of WSN Localization Problem

WSN node localization problem is formulated using thesingle hop range-based distribution technique to estimatethe position of the unknown node coordinates (X Y) withthe aid of anchor nodes (position of known nodes) co-ordinates (x y) Anchor nodes are provided with a GPSdevice so it has the capability of automatically determiningits position Most of the nodes in the WSN are not equippedwith GPS due to high cost To measure the coordinates of Nunknown nodes the procedure followed is given below

Step 1 Randomly initialize the N unknown nodes and Manchor nodes within the communication range (R) Anchornodes measure their position and communicate their co-ordinates to their neighbors For all iterations the nodewhich settles at the end is termed as the reference node andthis node will act as anchor node

Step 2 +ree or more anchor nodes within the commu-nication range of a node are considered as a localized node

Step 3 Let (x y) be the coordinates of the target node to bedetermined and di be the distance between the target nodeand the ith anchor node

di

xminus xi( 11138572

+ yminusyi( 11138572

1113969

(14)

Step 4 +e optimization problem is formulated to minimizethe error of the localization problem +e objective functionfor the localization problem is formulated as

(a) (b)

Figure 1 (a) Individual Salp (b) Salps chain [37]

6 Journal of Computer Networks and Communications

f(x y) min sumM

i1

xminus xi( )2 + yminusyi( )2

radic( )

2 (15)

whereM is the anchor nodes within the transmission range(R) of the target node

Step 5 All target localized nodes (NL) are determined thewhole localization error is calculated as the mean of thesquare of distances of the estimated coordinates node (xi yi)and the actual node coordinates (Xi Yi) for i 1 2 NL

EL 1NL

sumL

i1

xi minusXi( )2 + yi minusYi( )2

radic( ) (16)

e performance of SSA algorithm evaluated using ELand the number of nonlocalized nodes NNL whereNNL [NminusNL]

Step 6 Repeat the steps 2ndash5 until all unknowntarget nodesget localized or no more nodes can be localized

5 Experimental Analysis

In this section the proposed WSN localization approach isevaluated under dierent scenarios and its performance iscompared to four other swarm-based algorithms (PSOBOA FA and GWO) in terms of localization accuracy andcomputing time e computations of the dierent algo-rithms are performed using MATLAB R2012b on a machineof Intel Core i7 CPU 4GB RAM and Windows7 operatingsystem e parametersrsquo values of the deployment area areshown in Table 1

For BOA the sensory modality c is set 001 whereasthe initial value of power exponent a is set to 01 [25] ForPSO initial values of ω 07 and c1 c2 1494 were rec-ommended for faster convergence after experimental tests[25] For FA the randomization parameter α is set to 025the absorption coecient c is set to 10 and the initial

attractiveness parameter β is set to 1 For GWO the pa-rameter a linearly decreases in the interval of [2 to 0] and theC parameter linearly increases from 0 to 2 [6] Finally forSSA c1 is calculated using equation (2) whereas c2 07 andc3 03 and c2 c3 are random numbers uniformly generatedfrom the interval [0 1]

51 Sensor Nodes Localization Using SSA In all conductedexperiments the coordinates of central nodes (N) anddestination nodes (M) are randomly congured during theconstruction of the deployment area e deployment areaincludes three types of nodes anchor nodes whose knownposition target nodes whose unknown position and lo-calized nodes whose positions are already estimated In thissection the performance of the SSA-based localization al-gorithm is evaluated under dierent scenarios using dif-ferent numbers of target nodes and dierent numbers ofanchors as shown in Figure 3

52 Comparison among Dierent Localization AlgorithmsIn this section SSA and the other swarm algorithms havebeen evaluated under dierent scenarios in terms of local-ization error computing time and number of localizednodes e obtained results of the dierent algorithms areshown in Table 2

Under the dierent scenarios (number of nodesnumberof anchors) it is noticed that for all the localization algo-rithms increasing the number of iterations increases both ofthe number of localized nodes and the computing time whilereduces the localization error is notice is rational because

Initialize the Salp population xi (i = 1 2 hellip n) considering ub and lbWhile (end condition is not satisfied)Calculate the fitness of each search agent (salp)F = the best search agentUpdate c1 by equation (2)

Update the position of the leading Salp by equation (1)Else

Update the position of the follower Salp by equation (4)End

End

EndReturn F

Amend the Salps based on the upper and lower bounds of variables

If (i == 1)For each Salp (xi)

Figure 2 Pseudocode of the SSA algorithm

Table 1 Parameters setting of simulation environment

Parameters ValuesSensor nodes Varies on sum6

i1ilowast 25Anchor nodes Varies on increment i i + 5Node transmission range (R) 30mDeployment area 100mtimes 100mMaximum number of iterations 100

Journal of Computer Networks and Communications 7

increasing the number of iterations means higher amountsof computations which requires longer computation timeOn the other side increasing the number of iterations meansthat the chance to nd a better solution get bigger hence thenumber of localized nodes get larger and the value of lo-calization error get smaller For better results analysis theexperimental results are summarized in Table 3

Based on Table 3 regarding the mean localization error(EL) there is no clear pattern that can be detected to rep-resent the relationship between this performance metric andthe number of target nodes and anchor nodes However it isnoticed that SSA has the best results regarding this per-formance metric compared to PSO BOA FA and GWOparticularly when the numbers of target nodes and anchor

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(a)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(b)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(c)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(d)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(e)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(f )

FIGURE 3 Node localization using dierent numbers of target nodes and anchor nodes (a) Target nodes 25 anchor nodes 10 (b) Targetnodes 50 anchor nodes 15 (c) Target nodes 75 anchor nodes 20 (d) Target nodes 100 anchor nodes 25 (e) Target nodes 125anchor nodes 30 (f ) Target nodes 150 anchor nodes 35

8 Journal of Computer Networks and Communications

nodes are increased Regarding the computing time it isnoticed that increasing the number of target nodes andanchor nodes increase the computing time for all local-ization algorithms However once again SSA has the bestcomputing time compared to PSO BOA FA and GWOFinally regarding the number of localized nodes (NL) it isnoticed that SSA has the best results compared to otherlocalization algorithms In addition it is noticed that SSAhas inferior EL when the percentage of (number of anchornodesnumber of target nodes) becomes larger such as(1025) in the first case in Table 3 +at is because thelocalization accuracy increases when the anchor densityincreases with respect to the number of the target nodes[25]

+e graphical representations of the experimentalresults for the different performance metrics are shown inFigures 4ndash6

6 Conclusion

Accurate node localization concerns many applications thatadopt WSNs In this paper a node localization algorithm hasbeen proposed based on a novel bioinspired algorithm calledSalp Swarm Algorithm (SSA) which handled the node lo-calization problem as an optimization problem+e proposedalgorithm has been implemented and validated in differentWSN deployments using different numbers of target nodesand anchor nodes Moreover the proposed algorithm hasbeen evaluated and compared to four well-known optimi-zation algorithms namely PSO BOA FA and GWO interms of localization accuracy computing time and severallocalized nodes +e obtained simulation results have provedthe superiority of the proposed algorithm compared to theother localization algorithms regarding the different perfor-mance metrics In the future work the proposed approach

Table 2 Performance metrics of different localization algorithms

Targetnodes

Anchornodes

No ofiterations

PSO BOA FA GWO SSAEL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10

25 0818 040 16 0232 037 22 0265 06 19 0744 022 20 0465 035 2250 0812 040 15 0225 038 23 0262 12 20 0741 041 21 0462 035 2375 0803 041 18 0223 039 25 0258 15 21 0741 054 21 0458 036 23100 0792 041 18 0221 040 25 0251 18 19 0740 079 23 0451 037 24

50 15

25 0419 071 41 0338 084 47 0477 16 46 0690 042 44 0477 067 4350 0426 073 47 0332 085 48 0473 19 49 0688 063 45 0472 069 4775 0429 076 46 0326 086 49 0465 25 49 0686 081 46 0468 069 48100 0434 076 48 0323 086 49 0465 35 48 0682 098 48 0464 070 50

75 20

25 0735 131 73 0257 149 66 0519 29 71 0641 072 72 0519 090 6950 0728 132 74 0257 149 66 0513 38 72 0641 095 72 0513 092 7275 0728 133 74 0255 150 70 0504 47 73 0638 13 73 0504 095 73100 0724 135 75 0253 152 72 0503 52 73 0635 14 74 0503 096 75

100 25

25 0661 210 97 0355 240 97 0711 38 98 0611 11 95 0511 131 9850 0658 216 97 0355 244 99 0709 42 98 0606 15 97 0509 133 9875 0642 217 99 0333 247 100 0702 56 99 0602 18 98 0502 136 99100 0641 220 100 0331 250 100 0704 63 98 0602 21 98 0504 137 100

125 30

25 0754 487 120 0549 384 122 0829 27 122 0589 15 122 0529 167 12350 0748 486 121 0548 385 123 0824 45 123 0580 22 123 0524 168 12475 0750 489 122 0534 386 124 0822 59 125 0580 28 123 0522 170 125100 0752 495 125 0534 389 124 0822 65 124 0572 33 125 0522 172 125

150 35

25 0625 541 145 0766 588 147 0911 25 149 0559 28 148 0511 212 14950 0622 542 146 0765 561 148 0909 42 150 0547 36 149 0509 214 14975 0619 544 148 0763 564 149 0904 64 150 0523 43 150 0504 216 150100 0616 545 150 0763 569 149 0904 72 149 0523 48 150 0504 218 150

Table 3 Summary of experimental results of the different localization algorithms

Target nodes Anchor nodesPSO BOA FA GWO SSA

EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10 079 040 18 022 038 25 025 18 19 074 054 21 045 035 2450 15 043 076 46 032 086 49 042 25 49 069 081 46 046 069 4875 20 072 135 75 025 152 68 051 47 73 064 095 72 050 096 73100 25 065 216 100 035 245 100 070 53 98 060 21 98 051 135 100125 30 074 490 123 054 387 124 082 65 124 058 28 123 052 170 125150 35 062 543 149 076 565 149 090 72 149 052 43 150 050 215 150

Journal of Computer Networks and Communications 9

0

01

02

03

04M

ean

loca

lizat

ion

erro

r (m

)05

06

07

08

09

1

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

Figure 4 e localization error of the dierent localization algorithms in dierent WSN deployments

0

1

2

3

4

5

6

7

8

9

10

Com

putin

g tim

e (S)

(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35

Figure 5 e computing time of the dierent localization algorithms in dierent WSN deployments

0 20 40 60 80 100 120 140Number of localized nodes

(Num

ber o

f tar

get n

odes

ndashnu

mbe

r of

anch

ors)

SSAGWOFA

BOAPSO

25 ndash 10

50 ndash 15

75 ndash 20

100 ndash 25

125 ndash 30

150 ndash 35

Figure 6 e number of localized nodes of the dierent localization algorithms in dierent WSN deployments

10 Journal of Computer Networks and Communications

can be hybridized with other algorithm to reduce the local-ization error

Data Availability

No data were used to support this study Because our articlediscusses the problem of localization in WSNs our exper-imental results have been applied based on several networksrsquosize

Conflicts of Interest

+e authors declare that they have no conflicts of interest

References

[1] R V Kulkarni G K Venayagamoorthy and M X ChengldquoBio-inspired node localization in wireless sensor networksrdquoin Proceedings of IEEE International Conference in SystemsMan and Cybernetics (SMC) pp 205ndash210 San Antonio TXUSA October 2009

[2] D Lavanya and S K Udgata ldquoSwarm intelligence-basedlocalization in wireless sensor networksrdquo in InternationalWorkshop on Multi-Disciplinary Trends in Artificial In-telligence pp 317ndash328 Springer Berlin Germany 2011

[3] J Wang R K Ghosh and S K Das ldquoA survey on sensorlocalizationrdquo Journal of Control Aeory and Applicationsvol 8 no 1 pp 2ndash11 2010

[4] J Aspnes T Eren D K Goldenberg et al ldquoA theory ofnetwork localizationrdquo IEEE Transactions on Mobile Com-puting vol 5 no 12 pp 1663ndash1678 2006

[5] S Goyal and M S Patterh ldquoModified bat algorithm for lo-calization of wireless sensor networkrdquo Wireless PersonalCommunications vol 86 no 2 pp 657ndash670 2015

[6] M M Ahmed E H Houssein A E Hassanien A Taha andE Hassanien ldquoMaximizing lifetime of wireless sensor net-works based on whale optimization algorithmrdquo in In-ternational Conference on Advanced Intelligent Systems andInformatics pp 724ndash733 Springer Berlin Germany 2017

[7] K S Low H A Nguyen and H Guo ldquoA particle swarmoptimization approach for the localization of a wireless sensornetworkrdquo in Proceedings of IEEE International Symposium inIndustrial Electronics (ISIE) pp 1820ndash1825 VancouverCanada June 2008

[8] D Manjarres J Del Ser S Gil-Lopez M Vecchio I Landa-Torres and R Lopez-Valcarce ldquoOn the application of a hybridharmony search algorithm to node localization in anchor-based wireless sensor networksrdquo in Proceedings of IEEE In-ternational Conference in Intelligent Systems Design andApplications (ISDA) pp 1014ndash1019 Cordoba Spain No-vember 2011

[9] M Li W Xiong and Q Liang ldquoAn improved abc-based nodelocalization algorithm for wireless sensor networkrdquo in Pro-ceedings of IEEE International Conference in Wireless Com-munications Networking and Mobile Computing (WiCOM)pp 1ndash4 Barcelona Spain September 2012

[10] A Tamizharasi R Arthi and K Murugan ldquoBio-inspiredalgorithm for optimizing the localization of wireless sensornetworksrdquo in Proceedings of IEEE International Conference inComputing Communications and Networking Technologies(ICCCNT) pp 1ndash5 Tiruchengode India June 2013

[11] C Tang R Liu and J Ni ldquoA novel wireless sensor networklocalization approach localization based on plant growth

simulation algorithmrdquo Elektronika ir Elektrotechnika vol 19no 8 pp 97ndash100 2013

[12] O D Jegede and K Ferens ldquoA genetic algorithm for nodelocalization in wireless sensor networksrdquo in Proceedings ofInternational Conference on Genetic and EvolutionaryMethods (GEM) Ae Steering Committee of the World Con-gress in Computer Science Computer Engineering and AppliedComputing (WorldComp) pp 1ndash7 Las Vegas NV USA July2013

[13] S Goyal and M S Patterh ldquoWireless sensor network local-ization based on cuckoo search algorithmrdquo Wireless personalcommunications vol 79 no 1 pp 223ndash234 2014

[14] L Dan and W Xian-bin ldquoAn improved PSO algorithm fordistributed localization in wireless sensor networksrdquo In-ternational Journal of Distributed Sensor Networks vol 11no 7 Article ID 970272 2015

[15] R Krishnaprabha and A Gopakumar ldquoPerformance ofgravitational search algorithm in wireless sensor networklocalizationrdquo in Proceedings of IEEE National Conference inCommunication Signal Processing and Networking (NCCSN)pp 1ndash6 Palakkad India October 2014

[16] B Peng and L Li ldquoAn improved localization algorithm basedon genetic algorithm in wireless sensor networksrdquo CognitiveNeurodynamics vol 9 no 2 pp 249ndash256 2015

[17] V O Sai C S Shieh T T Nguyen Y C Lin M F Horngand Q D Le ldquoParallel firefly algorithm for localization al-gorithm in wireless sensor networkrdquo in Proceedings of IEEEInternational Conference in Robot Vision and Signal Pro-cessing (RVSP) pp 300ndash305 Kaohsiung China November2015

[18] S Sivakumar and R Venkatesan ldquoMeta-heuristic approachesfor minimizing error in localization of wireless sensor net-worksrdquo Applied Soft Computing vol 36 pp 506ndash518 2015

[19] Z Sun L Tao X Wang and Z Zhou ldquoLocalization algorithmin wireless sensor networks based on multiobjective particleswarm optimizationrdquo International Journal of DistributedSensor Networks vol 11 no 8 Article ID 716291 2015

[20] A Arsic M Tuba and M Jordanski ldquoFireworks algorithmapplied to wireless sensor networks localization problemrdquo inProceedings of IEEE Congress on Evolutionary Computation(CEC) pp 4038ndash4044 Vancouver Canada July 2016

[21] C S Shieh V O Sai Y C Lin T F Lee T T Nguyen andQ D Le ldquoImproved node localization for WSN using heu-ristic optimization approachesrdquo in Proceedings of IEEE In-ternational Conference in Networking and NetworkApplications (NaNA) pp 95ndash98 Hokkaido Japan July 2016

[22] J Cheng and L Xia ldquoAn effective Cuckoo search algorithm fornode localization in wireless sensor networkrdquo Sensors vol 16no 9 p 1390 2016

[23] P T Daely and S Y Shin ldquoRange based wireless node lo-calization using Dragonfly Algorithmrdquo in Proceedings of IEEEInternational Conference in Ubiquitous and Future Networks(ICUFN) pp 1012ndash1015 Vienna Austria July 2016

[24] T T Nguyen J S Pan S C Chu J F Roddick and T K DaoldquoOptimization localization in wireless sensor network basedon multi-objective firefly algorithmrdquo Journal of NetworkIntelligence vol 1 no 4 pp 130ndash138 2016

[25] S Arora and S Singh ldquoNode localization in wireless sensornetworks using butterfly optimization algorithmrdquo ArabianJournal for Science and Engineering vol 42 no 8pp 3325ndash3335 2017

[26] S R Sujatha and M Siddappa ldquoNode localization method forwireless sensor networks based on hybrid optimization ofparticle swarm optimization and differential evolutionrdquo IOSR

Journal of Computer Networks and Communications 11

Journal of Computer Engineering vol 19 no 2 pp 07ndash122017

[27] R Kaur and S Arora ldquonature inspired range based wirelesssensor node localization algorithmsrdquo International Journal ofInteractive Multimedia and Artificial Intelligence vol 4pp 7ndash17 2017

[28] A +arwat E H Houssein M M Ahmed A E Hassanienand T Gabel ldquoMOGOA algorithm for constrained and un-constrained multi-objective optimization problemsrdquo AppliedIntelligence vol 48 no 8 pp 2268ndash2283 2017

[29] A A Ewees M Abd Elaziz and E H Houssein ldquoImprovedgrasshopper optimization algorithm using opposition-basedlearningrdquo Expert Systems with Applications vol 112pp 156ndash172 2018

[30] A G Hussien E H Houssein and A E Hassanien ldquoA binarywhale optimization algorithm with hyperbolic tangent fitnessfunction for feature selectionrdquo in Proceedings of Eighth In-ternational Conference on Intelligent Computing and In-formation Systems (ICICIS) pp 166ndash172 Cairo EgyptDecember 2017

[31] B Xue M Zhang and W N Browne ldquoParticle swarm op-timization for feature selection in classification a multi-objective approachrdquo IEEE Transactions on Cyberneticsvol 43 no 6 pp 1656ndash1671 2013

[32] S Arora and S Singh ldquoButterfly algorithmwith levy flights forglobal optimizationrdquo in Proceedings of IEEE InternationalConference in Signal Processing Computing and Control(ISPCC) pp 220ndash224 Waknaghat India September 2015

[33] S Arora and S Singh ldquoAn improved butterfly optimizationalgorithm with chaosrdquo Journal of Intelligent and Fuzzy Sys-tems vol 32 no 1 pp 1079ndash1088 2017

[34] R B Blair and A E Launer ldquoButterfly diversity and humanland use species assemblages along an urban grandientrdquoBiological Conservation vol 80 no 1 pp 113ndash125 1997

[35] X S Yang and X He ldquoFirefly algorithm recent advances andapplicationsrdquo 2013 httpsarxivorgabs13083898

[36] X S Yang ldquoFirefly algorithms for multimodal optimizationrdquoin Proceedings of International Symposium on Stochastic Al-gorithms pp 169ndash178 Sapporo Japan October 2009

[37] S Mirjalili S M Mirjalili and A Lewis ldquoGrey wolf opti-mizerrdquo Advances in Engineering Software vol 69 pp 46ndash612014

[38] L P Madin ldquoAspects of jet propulsion in Salpsrdquo CanadianJournal of Zoology vol 68 no 4 pp 765ndash777 1990

[39] S Mirjalili A H Gandomi S Z Mirjalili S Saremi H Farisand S M Mirjalili ldquoSalp swarm algorithm a bio-inspiredoptimizer for engineering design problemsrdquo Advances inEngineering Software vol 114 pp 163ndash191 2017

[40] P A Anderson and Q Bone ldquoCommunication betweenindividuals in Salp chains II Physiologyrdquo Proceedings of theRoyal Society of London Series B Biological Sciences vol 210no 1181 pp 559ndash574 1980

[41] H T Ibrahim W J Mazher O N Ucan and O BayatldquoFeature selection using salp swarm algorithm for real bio-medical datasetsrdquo IJCSNS vol 17 no 12 pp 13ndash20 2017

12 Journal of Computer Networks and Communications

Dan and Xian-bin have presented a distributed two-phase PSO algorithm for efficiently and precisely localizingthe sensor nodes in addition to solving the flip ambiguityproblem In the first phase the initial search space is definedusing the bounding box method In the second phase arefinement process is performed for correcting the errorresulting from the flip ambiguity Additionally the proposedalgorithm attempted localizing sensor nodes that have onlytwo references or three near-collinear references +econducted simulation has proved the effectiveness of theproposed algorithm [14]

Krishnaprabha and Gopakumar have proposed a nodelocalization algorithm based on gravitational search algo-rithm (GSA) In the proposed work node localization inWSNs is formulated as a nonlinear optimization problemAlso the proposed algorithm tried to handle the flip am-biguity problem and to localize the sensor nodes that col-linear anchor nodes through the refinement phase +eobtained simulation results have shown that the proposedlocalization algorithm has good performance [15]

Peng and Li have focused on range-free localization as acost-effective alternative compared to range-based ap-proaches However they noticed that range-free localizationsuffers from higher localization error compared to the range-based algorithms In order to deal with this problem theypresented an improved version for DV-Hop which is apopular rang-free approach that depends on hop-distanceestimation +e improvement in the DV-Hop algorithm isperformed based on a genetic algorithm Simulation resultshave shown that the proposed localization algorithm hasbetter localization accuracy compared to other localizationalgorithms [16]

Sai et al have presented a hybrid node localizationalgorithm in WSNs which depends on the measurementsof the received signal strength (RSS) and parallel fireflyalgorithm (PFA) Taking into consideration the distancefactor the proposed algorithm transforms the node lo-calization problem in WSN into a nonlinear un-constrained optimization problem that is defined by anenhanced objective function In the proposed algorithmPFA estimates the coordinates of sensor nodes using thedistances between a sensor and a few numbers of its 1-hopneighbors Simulation results have shown that the pro-posed approach is better than PSO GA PFA and RSS interms of localization accuracy [17]

Sivakumar and Venkatesan have provided two-phasenode localization algorithm in WSNs In the first phase thepositions of sensor nodes are roughly estimated using arange-free localization method called mobile anchor po-sitioning (MAP) In the second phase the proposed al-gorithm attempts to reduce the localization error byemploying a certain metaheuristic approach In their workto perform the second phase they employed bat optimi-zation algorithm (BOA) modified cuckoo search (MCS)and firefly optimization algorithm (FOA) resulting in threelocalization algorithms namely BOA-MAP MCS-MAPand FOA-MAP +e experimental results have shownthat FOA-MAP is better than both BOA-MAP and MCS-MAP in terms of root mean square error (RMSE) [18]

Sun et al have proposed a multiobjective node locali-zation algorithm based on multiobjective particle swarmoptimization called multiobjective particle swarm optimi-zation localization algorithm (MOPSOLA) +e multi-objective functions include the space distance and thegeometric topology constraints In the proposed algorithmthe size of the archive remains limited using a dynamicmethod Simulation results have shown that the proposedalgorithm has achieved significant enhancements in terms oflocalization accuracy and convergence rate [19]

Arsic et al have proposed a node localization algorithmbased on fireworks algorithm (FWA) +e proposed al-gorithm provides optimal results in case of no rangingerrors and provides good results in case of ranging errorsMoreover they proposed an enhanced fireworks algorithm(EFWA) which achieved better results compared to FWAAlso the proposed localization algorithm outperforms theexisting PSO-based localization algorithm [20]

Shieh et al have compared several well-known heuristicssuch as genetic algorithm (GA) and particle swarm opti-mization (PSO) to more recent methods such as grey wolfoptimizer (GWO) firefly algorithm (FA) and brain stormoptimization (BSO) algorithms in terms of sensor nodeslocalization accuracy Also they proposed an enhancementin the localization algorithms to increase the number oflocalized nodes +e improved algorithms have been com-pared to the original ones in terms of the number of localizednodes and execution time in different network deployments[21]

Cheng and Xia have proposed a cuckoo search (CS)algorithm-based node localization algorithm In theproposed method the step size has been modified toobtain a global optimal solution in a short time Also thecandidate solutionsrsquo fitness is used for constructing mu-tation probability to avoid local convergence +e per-formance of the proposed algorithm has been evaluatedusing different anchor density node density and com-munication range in terms of average localization errorand localization success ratio +e simulation results haveproved that the proposed algorithm is better than thestandard CS and PSO in terms of average localizationerror and convergence time [22]

Daely and Shin have proposed a node localizationalgorithm based on dragonfly algorithm (DA) optimiza-tion algorithm +e proposed localization algorithm wasdesigned to determine the positions of the nodes whichare randomly distributed in a specific area +e simulationresults proved that the proposed DA based algorithm isbetter than PSO in terms of localization accuracy [23]

Nguyen et al have employed the multiobjective fireflyalgorithm to estimate the coordinates of sensor nodes inWSNs +e used objective functions depend on two criterianamely the nodesrsquo distance constraint and geometric to-pology constraint +e simulation results have shown thesuperiority of the proposed algorithm compared to well-known localization methods in terms of localization accu-racy and convergence rate [24]

Arora and Singh have used the butterfly optimizationalgorithm to localize the sensor nodes in WSNs +e

Journal of Computer Networks and Communications 3

proposed localization algorithm has been validated usingdifferent numbers of nodes with distance measurementscorrupted through the Gaussian noise +e simulation re-sults have shown that the proposed localization algorithm isbetter than several well-known localization algorithms in-cluding particle swarm optimization (PSO) and firefly al-gorithm (FA) in terms of localization accuracy [25]

Ahmed et al have presented a node localization algo-rithm based on whale optimization algorithm (WOA) whosemain objective is to localize sensor nodes in WSNs accu-rately [6] Also Sujatha and Siddappa have proposed ahybrid localization algorithm based on dynamic weightparticle swarm (DWPSO) and differential evolution (DE)algorithms +e authors observed that decreasing the squareerror of estimated and measured distance can improve thelocalization accuracy +e obtained simulation resultsproved that the proposed localization algorithm is betterthan the linearization method (LM) in terms of localizationaccuracy and performance stability [26]

Kaur and Arora have compared the performance ofseveral bioinspired algorithms including flower pollinationalgorithm (FPA) firefly algorithm (FA) grey wolf optimi-zation (GWO) and particle swarm optimization (PSO) inlocalizing the sensor nodes ofWSNs+e performance of thedifferent algorithms has been evaluated in terms of severalperformance metrics including localization accuracycomputing time and several localized nodes +e simulationresults have shown the superiority of FPA compared to theother algorithms in terms of localization accuracy [27]

To the best of our knowledge the Salp Swarm Algorithm(SSA) algorithm was never used for the localization probleminWSNs so far +erefore the main objective of this paper isto employ the SSA algorithm for handling the localizationproblem in WSNs and is to evaluate its performance againstseveral well-known swarm intelligence algorithms+e basicideas behind the SSA and other swarm algorithms are givenin the next section

3 Swarm Intelligence Algorithms

Swarm intelligence (SI) is a relatively new interdisciplinaryfield of research which has gained huge popularity in thesedays Algorithms belonging to this domain draw inspirationfrom the collective intelligence emerging from the behaviorof a group of social insects (like bees termites and wasps) Ithas successfully been applied to several real-world optimi-zation problems [28ndash30] In this section we review some ofthese algorithms that are employed in this paper in order tolocalize the nodes of WSNs

31 Particle Swarm Optimization Particle swarm optimi-zation (PSO) is a swarm optimization algorithm proposed byEberhart and Kennedy in 1995 It is inspired by the collectivebehavior of bird flocking and fish schooling It employsseveral particles that simulate a swarm population movingaround in the search space to find the best solution Eachparticle provides a candidate solution for the problem and isusually represented as a point in D-dimensional space +e

position of each particle is represented as a vectorxi (xi1 middot xi2 middot middot xi D) Particles move in the search spaceto find optimal solutions and each particle has a velocityrepresented as a vector vi (vi1 middot vi2 middot middot vi D) Duringmovement the position of each particle is updated based onthe best position that has been achieved by the particle itself(pbest) and the best position that has been achieved by all theparticles so far (gbest) +e velocity and the position of eachparticle are updated as shown below [31]

xt+1id x

tid + v

t+1id (1)

vt+1id wlowast v

tid + c1 lowast r1 lowast pid minusx

tid1113872 1113873 + c2 lowast r2 lowast pgd minusx

tid1113872 1113873

(2)

where t refers to the iteration number w is inertia weightthat aims at determining the effect of previous velocities oncurrent velocity c1 and c2 are acceleration constants r1 andr2 are random variables whose values are normally dis-tributed in [0 1] and pid and pgd indicate the elements ofpbest and gbest in the dth dimension respectively +e ve-locity is limited by a predefined maximum velocity vmax

32 Butterfly Optimization Algorithm Butterfly optimiza-tion algorithm (BOA) is a swarm intelligence algorithm thatis recently proposed by Arora and Singh [32] It is inspiredfrom the food-foraging behavior of butterflies Biologicallybutterflies find the source of food using sense receptors +erole of these sense receptors also called chemoreceptors isto sense fragrancesmell [33]

In BOA butterflies are the search agents that performoptimization Each butterfly emits fragrance with some in-tensity +is fragrance is propagated and sensed by otherbutterflies in the region +e fragrance emitted by butterfly iscorrelated with the butterflyrsquos fitness +is means that thefragrance of a butterfly changes according to its current lo-cation [32] When a butterfly is able to sense fragrance fromany other butterfly that is larger than its fragrance it will movetoward the latter and this phase is called global search On theother side when a butterfly cannot sense fragrance from otherbutterflies that is larger than its local fragrance it will moverandomly and this phase is called local search [34] In BOAthe fragrance is defined as a function of the physical intensityof stimulus [34] as shown below

fi cIa (3)

where fi is the perceived magnitude of fragrance c is thesensory modality I is the stimulus intensity and a is thepower exponent dependent on modality which accountsvarying degree of absorption As mentioned before there aretwo phases in the BOA algorithm namely local search phaseand global search phase In global search phase the butterflymoves one step toward the best butterflysolution glowast whichcan be formulated as

xt+1i x

ti + r

2times glowast minusx

ti1113872 1113873 times fi (4)

where xti is the solution vector xi of ith butterfly in iteration

number t glowast is the best solution found in the current stage

4 Journal of Computer Networks and Communications

fi is the fragrance of ith butterfly and r is a random numberin [0 1] Local search phase is formulated as

xt+1i x

ti + r

2times x

tk minusx

tj1113872 1113873 times fi (5)

where xtj and xt

k are jth and kth butterflies chosen randomlyfrom the solution space Equation (5) is considered a localrandom walk if and only if xt

j and xtk belongs to the same

subswarm and r is a random number in [0 1] Search forfood and mating partner by butterflies can occur at bothlocal and global scale therefore a switch probability p isemployed in BOA to switch between common global searchand intensive local search

33 Firefly Algorithm Firefly algorithm (FA) is a nature-inspired algorithm proposed by Yang and He [35] It mimicsthe social behaviors and flashing patterns of fireflies FAdepends on the next three idealized rules [36]

(i) Fireflies are unisex whichmeans that a firefly can getattracted to any other firefly regardless their sex

(ii) +e attractiveness of fireflies is directly proportionalto their brightness +us for any two flashingfireflies the firefly with less brightness moves to-ward the one with higher brightness If there is no abrighter firefly than a particular firefly the lattermoves randomly

(iii) +e brightness of a firefly is calculated using anobjective function

A fireflyrsquos attractiveness is proportional to the light in-tensity visualized by other fireflies in the region thereforethe relationship between the attractiveness β and the dis-tance r can be formulated as

β β0eminuscr2

(6)

where β0 is the brightness at distance r 0 and c is the lightabsorption coefficient +e movement of a firefly i toward amore attractive (brighter) firefly j is represented as

xt+1i x

ti + β0e

minuscr2ij x

tj minus x

ti1113872 1113873 + αtϵ

ti (7)

where rij is the distance between the fireflies i and j that iscalculated using the Euclidean norm the second term is dueto the attraction the third term is a randomization with αt

being the randomization parameter and ϵti is a vector thatincludes random numbers

34 Grey Wolf Optimizer Grey wolf optimizer (GWO) is arecent swarm intelligence algorithm inspired by the greywolf community It is developed by Mirjalili et al in 2014Grey wolf is a very dangerous creature which belongs to theCanidae family Grey wolves usually live in packs that consistof 5 to 12 wolves Each group has social dominance hier-archy alpha beta and omega in order +e alphas are amale and female which represent the leaders of the group+e betas are the second level of management hierarchy+eomegas are the final level in the hierarchy [37]

In order to mathematically model the social hierarchy ofwolves in GWO the fittest solution is referred to as the alpha(α) Consequently the second and third best solutions arebeta (szlig) and delta (δ) respectively +e rest of candidatesolutions are omega (ω) [35]+emathematical model of theencircling behavior is represented as follows [35]

Drarr

| Crarr

Xrarr

p (t)minus Xrarr

(t)|

Xrarr

(t + 1) |Xrarr

p(t)minus Ararr

Drarr

|(8)

where t indicates the current iteration Crarr

2 rrarr

2Ararr

2 ararr

rrarr

1 minus ararr X

rarrm is the position vector of the wolf r1

and r2 are random vectors in [0 1] and linearly varies from 2to 1 C and A are coefficient vectors and Xp is the positionvector of the prey During the optimization ωwolves updatetheir positions around αβ and δ +erefore the ω wolvescan reposition with respect to α β and δ as shown below[35]

Drarrα C

rarr1X

rarrαminus X

rarr11138681113868111386811138681113868

11138681113868111386811138681113868Drarrβ C

rarr2X

rarrβminus X

rarr11138681113868111386811138681113868

11138681113868111386811138681113868Drarrδ

Crarr

3Xrarrδ minus X

rarr11138681113868111386811138681113868

11138681113868111386811138681113868

Xrarr1 X

rarrαminus A

rarr1(D

rarrα)X

rarr2 X

rarrβminus A

rarr2(D

rarrβ)X

rarr3

Xrarrδ minus A

rarr3(D

rarrδ)

Xrarr

(t + 1) Xrarr

1 + Xrarr

2 + Xrarr

3

3

(9)

With these equations a search agent updates its positionaccording to α β and δ in an n-dimensional search space Inaddition the final position would be in a random placewithin a circle which is defined by the positions of αβ and δin the search space [37]

35 Salp SwarmAlgorithm Salps are part of Salpidae familywith the limpid cylinder design body +ey look like jelly-fishes in texture andmovement+e shape of a Salp is shownin Figure 1(a) +e water is pushed Salps bodies to moveforward [38] Generally the biological research about Salpsis still in its early stages because their living environmentsare hardly accessible and it is very difficult to keep them inlaboratory environment Salps swarming attitude is the maininspiration to build Salp swarm algorithm [39] Salpscompose a swarm in profound oceans which is called Salpchain +is chain is illustrated in Figure 1(b) +is chain canhelp to achieve better locomotion during the foragingprocess [40]

Originally the Salps population is divided into twogroups to formulate the mathematical model for Salp chainshead and followers +e head position is at the beginning ofthe chain while the rest of the chain is referred to as thefollowers [41]

+e Salps location is determined likewise swarm-basedmethods by an n-dimensional search area via considering

Journal of Computer Networks and Communications 5

the number of variables inside the presented problemrepresented by n Accordingly a two-dimensional matrixdescribed as x will reservation the position of all Salps It issupposed too in search space there is ldquoFrdquo which is a foodsource as the target of the swarm +e following equation issuggested to upgrade the leader location [37]

x1j

Fj + c1 ubj minus lbj1113872 1113873c2 + lbj1113872 1113873 c3 ge 0

Fj minus c1 ubj minus lbj1113872 1113873c2 + lbj1113872 1113873 c3 lt 0

⎧⎪⎨

⎪⎩(10)

where x1j shows the position of the first Salp (leader) in

the jth dimension Fj is the position of the food source in thejth dimension ubj indicates the upper bound of jth di-mension lbj indicates the lower bound of jth dimension andc1 cthe2 and c3 are random numbers

Equation (10) shows that the leader only updates itsposition with respect to the food source +e coefficient c1 isthe most important parameter in SSA because it balancesexploration and exploitation defined as follows [37]

c1 2eminus(4lL)2

(11)

where l is the current iteration and L is the maximumnumber of iterations

+e parameters c2 and c3 are random numbers uni-formly generated in the interval of [0 1] In fact they dictateif the next position in jth dimension should be towardspositive infinity or negative infinity as well as the step size Toupdate the position of the followers the following equationsis utilized (Newtonrsquos law of motion) [37]

xij

12

at2

+ v0t (12)

where ige 2 xij shows the position of ith follower Salp in jth

dimension t is time v0 is the initial speed and a vfinalv0where v x minus x0t

Because the time in optimization is an iteration thediscrepancy between iterations is equal to 1 and consideringv0 0 this equation can be expressed as follows [37]

xij

12

xij + x

iminus1j1113872 1113873 (13)

where ige 2 and xij shows the position of ith follower Salp in

jth dimension Using equations (1) and (4) the Salp chainscan be simulated Figure 2 shows the pseudocode to im-plement the SSA algorithm

4 Formulation of WSN Localization Problem

WSN node localization problem is formulated using thesingle hop range-based distribution technique to estimatethe position of the unknown node coordinates (X Y) withthe aid of anchor nodes (position of known nodes) co-ordinates (x y) Anchor nodes are provided with a GPSdevice so it has the capability of automatically determiningits position Most of the nodes in the WSN are not equippedwith GPS due to high cost To measure the coordinates of Nunknown nodes the procedure followed is given below

Step 1 Randomly initialize the N unknown nodes and Manchor nodes within the communication range (R) Anchornodes measure their position and communicate their co-ordinates to their neighbors For all iterations the nodewhich settles at the end is termed as the reference node andthis node will act as anchor node

Step 2 +ree or more anchor nodes within the commu-nication range of a node are considered as a localized node

Step 3 Let (x y) be the coordinates of the target node to bedetermined and di be the distance between the target nodeand the ith anchor node

di

xminus xi( 11138572

+ yminusyi( 11138572

1113969

(14)

Step 4 +e optimization problem is formulated to minimizethe error of the localization problem +e objective functionfor the localization problem is formulated as

(a) (b)

Figure 1 (a) Individual Salp (b) Salps chain [37]

6 Journal of Computer Networks and Communications

f(x y) min sumM

i1

xminus xi( )2 + yminusyi( )2

radic( )

2 (15)

whereM is the anchor nodes within the transmission range(R) of the target node

Step 5 All target localized nodes (NL) are determined thewhole localization error is calculated as the mean of thesquare of distances of the estimated coordinates node (xi yi)and the actual node coordinates (Xi Yi) for i 1 2 NL

EL 1NL

sumL

i1

xi minusXi( )2 + yi minusYi( )2

radic( ) (16)

e performance of SSA algorithm evaluated using ELand the number of nonlocalized nodes NNL whereNNL [NminusNL]

Step 6 Repeat the steps 2ndash5 until all unknowntarget nodesget localized or no more nodes can be localized

5 Experimental Analysis

In this section the proposed WSN localization approach isevaluated under dierent scenarios and its performance iscompared to four other swarm-based algorithms (PSOBOA FA and GWO) in terms of localization accuracy andcomputing time e computations of the dierent algo-rithms are performed using MATLAB R2012b on a machineof Intel Core i7 CPU 4GB RAM and Windows7 operatingsystem e parametersrsquo values of the deployment area areshown in Table 1

For BOA the sensory modality c is set 001 whereasthe initial value of power exponent a is set to 01 [25] ForPSO initial values of ω 07 and c1 c2 1494 were rec-ommended for faster convergence after experimental tests[25] For FA the randomization parameter α is set to 025the absorption coecient c is set to 10 and the initial

attractiveness parameter β is set to 1 For GWO the pa-rameter a linearly decreases in the interval of [2 to 0] and theC parameter linearly increases from 0 to 2 [6] Finally forSSA c1 is calculated using equation (2) whereas c2 07 andc3 03 and c2 c3 are random numbers uniformly generatedfrom the interval [0 1]

51 Sensor Nodes Localization Using SSA In all conductedexperiments the coordinates of central nodes (N) anddestination nodes (M) are randomly congured during theconstruction of the deployment area e deployment areaincludes three types of nodes anchor nodes whose knownposition target nodes whose unknown position and lo-calized nodes whose positions are already estimated In thissection the performance of the SSA-based localization al-gorithm is evaluated under dierent scenarios using dif-ferent numbers of target nodes and dierent numbers ofanchors as shown in Figure 3

52 Comparison among Dierent Localization AlgorithmsIn this section SSA and the other swarm algorithms havebeen evaluated under dierent scenarios in terms of local-ization error computing time and number of localizednodes e obtained results of the dierent algorithms areshown in Table 2

Under the dierent scenarios (number of nodesnumberof anchors) it is noticed that for all the localization algo-rithms increasing the number of iterations increases both ofthe number of localized nodes and the computing time whilereduces the localization error is notice is rational because

Initialize the Salp population xi (i = 1 2 hellip n) considering ub and lbWhile (end condition is not satisfied)Calculate the fitness of each search agent (salp)F = the best search agentUpdate c1 by equation (2)

Update the position of the leading Salp by equation (1)Else

Update the position of the follower Salp by equation (4)End

End

EndReturn F

Amend the Salps based on the upper and lower bounds of variables

If (i == 1)For each Salp (xi)

Figure 2 Pseudocode of the SSA algorithm

Table 1 Parameters setting of simulation environment

Parameters ValuesSensor nodes Varies on sum6

i1ilowast 25Anchor nodes Varies on increment i i + 5Node transmission range (R) 30mDeployment area 100mtimes 100mMaximum number of iterations 100

Journal of Computer Networks and Communications 7

increasing the number of iterations means higher amountsof computations which requires longer computation timeOn the other side increasing the number of iterations meansthat the chance to nd a better solution get bigger hence thenumber of localized nodes get larger and the value of lo-calization error get smaller For better results analysis theexperimental results are summarized in Table 3

Based on Table 3 regarding the mean localization error(EL) there is no clear pattern that can be detected to rep-resent the relationship between this performance metric andthe number of target nodes and anchor nodes However it isnoticed that SSA has the best results regarding this per-formance metric compared to PSO BOA FA and GWOparticularly when the numbers of target nodes and anchor

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(a)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(b)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(c)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(d)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(e)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(f )

FIGURE 3 Node localization using dierent numbers of target nodes and anchor nodes (a) Target nodes 25 anchor nodes 10 (b) Targetnodes 50 anchor nodes 15 (c) Target nodes 75 anchor nodes 20 (d) Target nodes 100 anchor nodes 25 (e) Target nodes 125anchor nodes 30 (f ) Target nodes 150 anchor nodes 35

8 Journal of Computer Networks and Communications

nodes are increased Regarding the computing time it isnoticed that increasing the number of target nodes andanchor nodes increase the computing time for all local-ization algorithms However once again SSA has the bestcomputing time compared to PSO BOA FA and GWOFinally regarding the number of localized nodes (NL) it isnoticed that SSA has the best results compared to otherlocalization algorithms In addition it is noticed that SSAhas inferior EL when the percentage of (number of anchornodesnumber of target nodes) becomes larger such as(1025) in the first case in Table 3 +at is because thelocalization accuracy increases when the anchor densityincreases with respect to the number of the target nodes[25]

+e graphical representations of the experimentalresults for the different performance metrics are shown inFigures 4ndash6

6 Conclusion

Accurate node localization concerns many applications thatadopt WSNs In this paper a node localization algorithm hasbeen proposed based on a novel bioinspired algorithm calledSalp Swarm Algorithm (SSA) which handled the node lo-calization problem as an optimization problem+e proposedalgorithm has been implemented and validated in differentWSN deployments using different numbers of target nodesand anchor nodes Moreover the proposed algorithm hasbeen evaluated and compared to four well-known optimi-zation algorithms namely PSO BOA FA and GWO interms of localization accuracy computing time and severallocalized nodes +e obtained simulation results have provedthe superiority of the proposed algorithm compared to theother localization algorithms regarding the different perfor-mance metrics In the future work the proposed approach

Table 2 Performance metrics of different localization algorithms

Targetnodes

Anchornodes

No ofiterations

PSO BOA FA GWO SSAEL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10

25 0818 040 16 0232 037 22 0265 06 19 0744 022 20 0465 035 2250 0812 040 15 0225 038 23 0262 12 20 0741 041 21 0462 035 2375 0803 041 18 0223 039 25 0258 15 21 0741 054 21 0458 036 23100 0792 041 18 0221 040 25 0251 18 19 0740 079 23 0451 037 24

50 15

25 0419 071 41 0338 084 47 0477 16 46 0690 042 44 0477 067 4350 0426 073 47 0332 085 48 0473 19 49 0688 063 45 0472 069 4775 0429 076 46 0326 086 49 0465 25 49 0686 081 46 0468 069 48100 0434 076 48 0323 086 49 0465 35 48 0682 098 48 0464 070 50

75 20

25 0735 131 73 0257 149 66 0519 29 71 0641 072 72 0519 090 6950 0728 132 74 0257 149 66 0513 38 72 0641 095 72 0513 092 7275 0728 133 74 0255 150 70 0504 47 73 0638 13 73 0504 095 73100 0724 135 75 0253 152 72 0503 52 73 0635 14 74 0503 096 75

100 25

25 0661 210 97 0355 240 97 0711 38 98 0611 11 95 0511 131 9850 0658 216 97 0355 244 99 0709 42 98 0606 15 97 0509 133 9875 0642 217 99 0333 247 100 0702 56 99 0602 18 98 0502 136 99100 0641 220 100 0331 250 100 0704 63 98 0602 21 98 0504 137 100

125 30

25 0754 487 120 0549 384 122 0829 27 122 0589 15 122 0529 167 12350 0748 486 121 0548 385 123 0824 45 123 0580 22 123 0524 168 12475 0750 489 122 0534 386 124 0822 59 125 0580 28 123 0522 170 125100 0752 495 125 0534 389 124 0822 65 124 0572 33 125 0522 172 125

150 35

25 0625 541 145 0766 588 147 0911 25 149 0559 28 148 0511 212 14950 0622 542 146 0765 561 148 0909 42 150 0547 36 149 0509 214 14975 0619 544 148 0763 564 149 0904 64 150 0523 43 150 0504 216 150100 0616 545 150 0763 569 149 0904 72 149 0523 48 150 0504 218 150

Table 3 Summary of experimental results of the different localization algorithms

Target nodes Anchor nodesPSO BOA FA GWO SSA

EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10 079 040 18 022 038 25 025 18 19 074 054 21 045 035 2450 15 043 076 46 032 086 49 042 25 49 069 081 46 046 069 4875 20 072 135 75 025 152 68 051 47 73 064 095 72 050 096 73100 25 065 216 100 035 245 100 070 53 98 060 21 98 051 135 100125 30 074 490 123 054 387 124 082 65 124 058 28 123 052 170 125150 35 062 543 149 076 565 149 090 72 149 052 43 150 050 215 150

Journal of Computer Networks and Communications 9

0

01

02

03

04M

ean

loca

lizat

ion

erro

r (m

)05

06

07

08

09

1

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

Figure 4 e localization error of the dierent localization algorithms in dierent WSN deployments

0

1

2

3

4

5

6

7

8

9

10

Com

putin

g tim

e (S)

(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35

Figure 5 e computing time of the dierent localization algorithms in dierent WSN deployments

0 20 40 60 80 100 120 140Number of localized nodes

(Num

ber o

f tar

get n

odes

ndashnu

mbe

r of

anch

ors)

SSAGWOFA

BOAPSO

25 ndash 10

50 ndash 15

75 ndash 20

100 ndash 25

125 ndash 30

150 ndash 35

Figure 6 e number of localized nodes of the dierent localization algorithms in dierent WSN deployments

10 Journal of Computer Networks and Communications

can be hybridized with other algorithm to reduce the local-ization error

Data Availability

No data were used to support this study Because our articlediscusses the problem of localization in WSNs our exper-imental results have been applied based on several networksrsquosize

Conflicts of Interest

+e authors declare that they have no conflicts of interest

References

[1] R V Kulkarni G K Venayagamoorthy and M X ChengldquoBio-inspired node localization in wireless sensor networksrdquoin Proceedings of IEEE International Conference in SystemsMan and Cybernetics (SMC) pp 205ndash210 San Antonio TXUSA October 2009

[2] D Lavanya and S K Udgata ldquoSwarm intelligence-basedlocalization in wireless sensor networksrdquo in InternationalWorkshop on Multi-Disciplinary Trends in Artificial In-telligence pp 317ndash328 Springer Berlin Germany 2011

[3] J Wang R K Ghosh and S K Das ldquoA survey on sensorlocalizationrdquo Journal of Control Aeory and Applicationsvol 8 no 1 pp 2ndash11 2010

[4] J Aspnes T Eren D K Goldenberg et al ldquoA theory ofnetwork localizationrdquo IEEE Transactions on Mobile Com-puting vol 5 no 12 pp 1663ndash1678 2006

[5] S Goyal and M S Patterh ldquoModified bat algorithm for lo-calization of wireless sensor networkrdquo Wireless PersonalCommunications vol 86 no 2 pp 657ndash670 2015

[6] M M Ahmed E H Houssein A E Hassanien A Taha andE Hassanien ldquoMaximizing lifetime of wireless sensor net-works based on whale optimization algorithmrdquo in In-ternational Conference on Advanced Intelligent Systems andInformatics pp 724ndash733 Springer Berlin Germany 2017

[7] K S Low H A Nguyen and H Guo ldquoA particle swarmoptimization approach for the localization of a wireless sensornetworkrdquo in Proceedings of IEEE International Symposium inIndustrial Electronics (ISIE) pp 1820ndash1825 VancouverCanada June 2008

[8] D Manjarres J Del Ser S Gil-Lopez M Vecchio I Landa-Torres and R Lopez-Valcarce ldquoOn the application of a hybridharmony search algorithm to node localization in anchor-based wireless sensor networksrdquo in Proceedings of IEEE In-ternational Conference in Intelligent Systems Design andApplications (ISDA) pp 1014ndash1019 Cordoba Spain No-vember 2011

[9] M Li W Xiong and Q Liang ldquoAn improved abc-based nodelocalization algorithm for wireless sensor networkrdquo in Pro-ceedings of IEEE International Conference in Wireless Com-munications Networking and Mobile Computing (WiCOM)pp 1ndash4 Barcelona Spain September 2012

[10] A Tamizharasi R Arthi and K Murugan ldquoBio-inspiredalgorithm for optimizing the localization of wireless sensornetworksrdquo in Proceedings of IEEE International Conference inComputing Communications and Networking Technologies(ICCCNT) pp 1ndash5 Tiruchengode India June 2013

[11] C Tang R Liu and J Ni ldquoA novel wireless sensor networklocalization approach localization based on plant growth

simulation algorithmrdquo Elektronika ir Elektrotechnika vol 19no 8 pp 97ndash100 2013

[12] O D Jegede and K Ferens ldquoA genetic algorithm for nodelocalization in wireless sensor networksrdquo in Proceedings ofInternational Conference on Genetic and EvolutionaryMethods (GEM) Ae Steering Committee of the World Con-gress in Computer Science Computer Engineering and AppliedComputing (WorldComp) pp 1ndash7 Las Vegas NV USA July2013

[13] S Goyal and M S Patterh ldquoWireless sensor network local-ization based on cuckoo search algorithmrdquo Wireless personalcommunications vol 79 no 1 pp 223ndash234 2014

[14] L Dan and W Xian-bin ldquoAn improved PSO algorithm fordistributed localization in wireless sensor networksrdquo In-ternational Journal of Distributed Sensor Networks vol 11no 7 Article ID 970272 2015

[15] R Krishnaprabha and A Gopakumar ldquoPerformance ofgravitational search algorithm in wireless sensor networklocalizationrdquo in Proceedings of IEEE National Conference inCommunication Signal Processing and Networking (NCCSN)pp 1ndash6 Palakkad India October 2014

[16] B Peng and L Li ldquoAn improved localization algorithm basedon genetic algorithm in wireless sensor networksrdquo CognitiveNeurodynamics vol 9 no 2 pp 249ndash256 2015

[17] V O Sai C S Shieh T T Nguyen Y C Lin M F Horngand Q D Le ldquoParallel firefly algorithm for localization al-gorithm in wireless sensor networkrdquo in Proceedings of IEEEInternational Conference in Robot Vision and Signal Pro-cessing (RVSP) pp 300ndash305 Kaohsiung China November2015

[18] S Sivakumar and R Venkatesan ldquoMeta-heuristic approachesfor minimizing error in localization of wireless sensor net-worksrdquo Applied Soft Computing vol 36 pp 506ndash518 2015

[19] Z Sun L Tao X Wang and Z Zhou ldquoLocalization algorithmin wireless sensor networks based on multiobjective particleswarm optimizationrdquo International Journal of DistributedSensor Networks vol 11 no 8 Article ID 716291 2015

[20] A Arsic M Tuba and M Jordanski ldquoFireworks algorithmapplied to wireless sensor networks localization problemrdquo inProceedings of IEEE Congress on Evolutionary Computation(CEC) pp 4038ndash4044 Vancouver Canada July 2016

[21] C S Shieh V O Sai Y C Lin T F Lee T T Nguyen andQ D Le ldquoImproved node localization for WSN using heu-ristic optimization approachesrdquo in Proceedings of IEEE In-ternational Conference in Networking and NetworkApplications (NaNA) pp 95ndash98 Hokkaido Japan July 2016

[22] J Cheng and L Xia ldquoAn effective Cuckoo search algorithm fornode localization in wireless sensor networkrdquo Sensors vol 16no 9 p 1390 2016

[23] P T Daely and S Y Shin ldquoRange based wireless node lo-calization using Dragonfly Algorithmrdquo in Proceedings of IEEEInternational Conference in Ubiquitous and Future Networks(ICUFN) pp 1012ndash1015 Vienna Austria July 2016

[24] T T Nguyen J S Pan S C Chu J F Roddick and T K DaoldquoOptimization localization in wireless sensor network basedon multi-objective firefly algorithmrdquo Journal of NetworkIntelligence vol 1 no 4 pp 130ndash138 2016

[25] S Arora and S Singh ldquoNode localization in wireless sensornetworks using butterfly optimization algorithmrdquo ArabianJournal for Science and Engineering vol 42 no 8pp 3325ndash3335 2017

[26] S R Sujatha and M Siddappa ldquoNode localization method forwireless sensor networks based on hybrid optimization ofparticle swarm optimization and differential evolutionrdquo IOSR

Journal of Computer Networks and Communications 11

Journal of Computer Engineering vol 19 no 2 pp 07ndash122017

[27] R Kaur and S Arora ldquonature inspired range based wirelesssensor node localization algorithmsrdquo International Journal ofInteractive Multimedia and Artificial Intelligence vol 4pp 7ndash17 2017

[28] A +arwat E H Houssein M M Ahmed A E Hassanienand T Gabel ldquoMOGOA algorithm for constrained and un-constrained multi-objective optimization problemsrdquo AppliedIntelligence vol 48 no 8 pp 2268ndash2283 2017

[29] A A Ewees M Abd Elaziz and E H Houssein ldquoImprovedgrasshopper optimization algorithm using opposition-basedlearningrdquo Expert Systems with Applications vol 112pp 156ndash172 2018

[30] A G Hussien E H Houssein and A E Hassanien ldquoA binarywhale optimization algorithm with hyperbolic tangent fitnessfunction for feature selectionrdquo in Proceedings of Eighth In-ternational Conference on Intelligent Computing and In-formation Systems (ICICIS) pp 166ndash172 Cairo EgyptDecember 2017

[31] B Xue M Zhang and W N Browne ldquoParticle swarm op-timization for feature selection in classification a multi-objective approachrdquo IEEE Transactions on Cyberneticsvol 43 no 6 pp 1656ndash1671 2013

[32] S Arora and S Singh ldquoButterfly algorithmwith levy flights forglobal optimizationrdquo in Proceedings of IEEE InternationalConference in Signal Processing Computing and Control(ISPCC) pp 220ndash224 Waknaghat India September 2015

[33] S Arora and S Singh ldquoAn improved butterfly optimizationalgorithm with chaosrdquo Journal of Intelligent and Fuzzy Sys-tems vol 32 no 1 pp 1079ndash1088 2017

[34] R B Blair and A E Launer ldquoButterfly diversity and humanland use species assemblages along an urban grandientrdquoBiological Conservation vol 80 no 1 pp 113ndash125 1997

[35] X S Yang and X He ldquoFirefly algorithm recent advances andapplicationsrdquo 2013 httpsarxivorgabs13083898

[36] X S Yang ldquoFirefly algorithms for multimodal optimizationrdquoin Proceedings of International Symposium on Stochastic Al-gorithms pp 169ndash178 Sapporo Japan October 2009

[37] S Mirjalili S M Mirjalili and A Lewis ldquoGrey wolf opti-mizerrdquo Advances in Engineering Software vol 69 pp 46ndash612014

[38] L P Madin ldquoAspects of jet propulsion in Salpsrdquo CanadianJournal of Zoology vol 68 no 4 pp 765ndash777 1990

[39] S Mirjalili A H Gandomi S Z Mirjalili S Saremi H Farisand S M Mirjalili ldquoSalp swarm algorithm a bio-inspiredoptimizer for engineering design problemsrdquo Advances inEngineering Software vol 114 pp 163ndash191 2017

[40] P A Anderson and Q Bone ldquoCommunication betweenindividuals in Salp chains II Physiologyrdquo Proceedings of theRoyal Society of London Series B Biological Sciences vol 210no 1181 pp 559ndash574 1980

[41] H T Ibrahim W J Mazher O N Ucan and O BayatldquoFeature selection using salp swarm algorithm for real bio-medical datasetsrdquo IJCSNS vol 17 no 12 pp 13ndash20 2017

12 Journal of Computer Networks and Communications

proposed localization algorithm has been validated usingdifferent numbers of nodes with distance measurementscorrupted through the Gaussian noise +e simulation re-sults have shown that the proposed localization algorithm isbetter than several well-known localization algorithms in-cluding particle swarm optimization (PSO) and firefly al-gorithm (FA) in terms of localization accuracy [25]

Ahmed et al have presented a node localization algo-rithm based on whale optimization algorithm (WOA) whosemain objective is to localize sensor nodes in WSNs accu-rately [6] Also Sujatha and Siddappa have proposed ahybrid localization algorithm based on dynamic weightparticle swarm (DWPSO) and differential evolution (DE)algorithms +e authors observed that decreasing the squareerror of estimated and measured distance can improve thelocalization accuracy +e obtained simulation resultsproved that the proposed localization algorithm is betterthan the linearization method (LM) in terms of localizationaccuracy and performance stability [26]

Kaur and Arora have compared the performance ofseveral bioinspired algorithms including flower pollinationalgorithm (FPA) firefly algorithm (FA) grey wolf optimi-zation (GWO) and particle swarm optimization (PSO) inlocalizing the sensor nodes ofWSNs+e performance of thedifferent algorithms has been evaluated in terms of severalperformance metrics including localization accuracycomputing time and several localized nodes +e simulationresults have shown the superiority of FPA compared to theother algorithms in terms of localization accuracy [27]

To the best of our knowledge the Salp Swarm Algorithm(SSA) algorithm was never used for the localization probleminWSNs so far +erefore the main objective of this paper isto employ the SSA algorithm for handling the localizationproblem in WSNs and is to evaluate its performance againstseveral well-known swarm intelligence algorithms+e basicideas behind the SSA and other swarm algorithms are givenin the next section

3 Swarm Intelligence Algorithms

Swarm intelligence (SI) is a relatively new interdisciplinaryfield of research which has gained huge popularity in thesedays Algorithms belonging to this domain draw inspirationfrom the collective intelligence emerging from the behaviorof a group of social insects (like bees termites and wasps) Ithas successfully been applied to several real-world optimi-zation problems [28ndash30] In this section we review some ofthese algorithms that are employed in this paper in order tolocalize the nodes of WSNs

31 Particle Swarm Optimization Particle swarm optimi-zation (PSO) is a swarm optimization algorithm proposed byEberhart and Kennedy in 1995 It is inspired by the collectivebehavior of bird flocking and fish schooling It employsseveral particles that simulate a swarm population movingaround in the search space to find the best solution Eachparticle provides a candidate solution for the problem and isusually represented as a point in D-dimensional space +e

position of each particle is represented as a vectorxi (xi1 middot xi2 middot middot xi D) Particles move in the search spaceto find optimal solutions and each particle has a velocityrepresented as a vector vi (vi1 middot vi2 middot middot vi D) Duringmovement the position of each particle is updated based onthe best position that has been achieved by the particle itself(pbest) and the best position that has been achieved by all theparticles so far (gbest) +e velocity and the position of eachparticle are updated as shown below [31]

xt+1id x

tid + v

t+1id (1)

vt+1id wlowast v

tid + c1 lowast r1 lowast pid minusx

tid1113872 1113873 + c2 lowast r2 lowast pgd minusx

tid1113872 1113873

(2)

where t refers to the iteration number w is inertia weightthat aims at determining the effect of previous velocities oncurrent velocity c1 and c2 are acceleration constants r1 andr2 are random variables whose values are normally dis-tributed in [0 1] and pid and pgd indicate the elements ofpbest and gbest in the dth dimension respectively +e ve-locity is limited by a predefined maximum velocity vmax

32 Butterfly Optimization Algorithm Butterfly optimiza-tion algorithm (BOA) is a swarm intelligence algorithm thatis recently proposed by Arora and Singh [32] It is inspiredfrom the food-foraging behavior of butterflies Biologicallybutterflies find the source of food using sense receptors +erole of these sense receptors also called chemoreceptors isto sense fragrancesmell [33]

In BOA butterflies are the search agents that performoptimization Each butterfly emits fragrance with some in-tensity +is fragrance is propagated and sensed by otherbutterflies in the region +e fragrance emitted by butterfly iscorrelated with the butterflyrsquos fitness +is means that thefragrance of a butterfly changes according to its current lo-cation [32] When a butterfly is able to sense fragrance fromany other butterfly that is larger than its fragrance it will movetoward the latter and this phase is called global search On theother side when a butterfly cannot sense fragrance from otherbutterflies that is larger than its local fragrance it will moverandomly and this phase is called local search [34] In BOAthe fragrance is defined as a function of the physical intensityof stimulus [34] as shown below

fi cIa (3)

where fi is the perceived magnitude of fragrance c is thesensory modality I is the stimulus intensity and a is thepower exponent dependent on modality which accountsvarying degree of absorption As mentioned before there aretwo phases in the BOA algorithm namely local search phaseand global search phase In global search phase the butterflymoves one step toward the best butterflysolution glowast whichcan be formulated as

xt+1i x

ti + r

2times glowast minusx

ti1113872 1113873 times fi (4)

where xti is the solution vector xi of ith butterfly in iteration

number t glowast is the best solution found in the current stage

4 Journal of Computer Networks and Communications

fi is the fragrance of ith butterfly and r is a random numberin [0 1] Local search phase is formulated as

xt+1i x

ti + r

2times x

tk minusx

tj1113872 1113873 times fi (5)

where xtj and xt

k are jth and kth butterflies chosen randomlyfrom the solution space Equation (5) is considered a localrandom walk if and only if xt

j and xtk belongs to the same

subswarm and r is a random number in [0 1] Search forfood and mating partner by butterflies can occur at bothlocal and global scale therefore a switch probability p isemployed in BOA to switch between common global searchand intensive local search

33 Firefly Algorithm Firefly algorithm (FA) is a nature-inspired algorithm proposed by Yang and He [35] It mimicsthe social behaviors and flashing patterns of fireflies FAdepends on the next three idealized rules [36]

(i) Fireflies are unisex whichmeans that a firefly can getattracted to any other firefly regardless their sex

(ii) +e attractiveness of fireflies is directly proportionalto their brightness +us for any two flashingfireflies the firefly with less brightness moves to-ward the one with higher brightness If there is no abrighter firefly than a particular firefly the lattermoves randomly

(iii) +e brightness of a firefly is calculated using anobjective function

A fireflyrsquos attractiveness is proportional to the light in-tensity visualized by other fireflies in the region thereforethe relationship between the attractiveness β and the dis-tance r can be formulated as

β β0eminuscr2

(6)

where β0 is the brightness at distance r 0 and c is the lightabsorption coefficient +e movement of a firefly i toward amore attractive (brighter) firefly j is represented as

xt+1i x

ti + β0e

minuscr2ij x

tj minus x

ti1113872 1113873 + αtϵ

ti (7)

where rij is the distance between the fireflies i and j that iscalculated using the Euclidean norm the second term is dueto the attraction the third term is a randomization with αt

being the randomization parameter and ϵti is a vector thatincludes random numbers

34 Grey Wolf Optimizer Grey wolf optimizer (GWO) is arecent swarm intelligence algorithm inspired by the greywolf community It is developed by Mirjalili et al in 2014Grey wolf is a very dangerous creature which belongs to theCanidae family Grey wolves usually live in packs that consistof 5 to 12 wolves Each group has social dominance hier-archy alpha beta and omega in order +e alphas are amale and female which represent the leaders of the group+e betas are the second level of management hierarchy+eomegas are the final level in the hierarchy [37]

In order to mathematically model the social hierarchy ofwolves in GWO the fittest solution is referred to as the alpha(α) Consequently the second and third best solutions arebeta (szlig) and delta (δ) respectively +e rest of candidatesolutions are omega (ω) [35]+emathematical model of theencircling behavior is represented as follows [35]

Drarr

| Crarr

Xrarr

p (t)minus Xrarr

(t)|

Xrarr

(t + 1) |Xrarr

p(t)minus Ararr

Drarr

|(8)

where t indicates the current iteration Crarr

2 rrarr

2Ararr

2 ararr

rrarr

1 minus ararr X

rarrm is the position vector of the wolf r1

and r2 are random vectors in [0 1] and linearly varies from 2to 1 C and A are coefficient vectors and Xp is the positionvector of the prey During the optimization ωwolves updatetheir positions around αβ and δ +erefore the ω wolvescan reposition with respect to α β and δ as shown below[35]

Drarrα C

rarr1X

rarrαminus X

rarr11138681113868111386811138681113868

11138681113868111386811138681113868Drarrβ C

rarr2X

rarrβminus X

rarr11138681113868111386811138681113868

11138681113868111386811138681113868Drarrδ

Crarr

3Xrarrδ minus X

rarr11138681113868111386811138681113868

11138681113868111386811138681113868

Xrarr1 X

rarrαminus A

rarr1(D

rarrα)X

rarr2 X

rarrβminus A

rarr2(D

rarrβ)X

rarr3

Xrarrδ minus A

rarr3(D

rarrδ)

Xrarr

(t + 1) Xrarr

1 + Xrarr

2 + Xrarr

3

3

(9)

With these equations a search agent updates its positionaccording to α β and δ in an n-dimensional search space Inaddition the final position would be in a random placewithin a circle which is defined by the positions of αβ and δin the search space [37]

35 Salp SwarmAlgorithm Salps are part of Salpidae familywith the limpid cylinder design body +ey look like jelly-fishes in texture andmovement+e shape of a Salp is shownin Figure 1(a) +e water is pushed Salps bodies to moveforward [38] Generally the biological research about Salpsis still in its early stages because their living environmentsare hardly accessible and it is very difficult to keep them inlaboratory environment Salps swarming attitude is the maininspiration to build Salp swarm algorithm [39] Salpscompose a swarm in profound oceans which is called Salpchain +is chain is illustrated in Figure 1(b) +is chain canhelp to achieve better locomotion during the foragingprocess [40]

Originally the Salps population is divided into twogroups to formulate the mathematical model for Salp chainshead and followers +e head position is at the beginning ofthe chain while the rest of the chain is referred to as thefollowers [41]

+e Salps location is determined likewise swarm-basedmethods by an n-dimensional search area via considering

Journal of Computer Networks and Communications 5

the number of variables inside the presented problemrepresented by n Accordingly a two-dimensional matrixdescribed as x will reservation the position of all Salps It issupposed too in search space there is ldquoFrdquo which is a foodsource as the target of the swarm +e following equation issuggested to upgrade the leader location [37]

x1j

Fj + c1 ubj minus lbj1113872 1113873c2 + lbj1113872 1113873 c3 ge 0

Fj minus c1 ubj minus lbj1113872 1113873c2 + lbj1113872 1113873 c3 lt 0

⎧⎪⎨

⎪⎩(10)

where x1j shows the position of the first Salp (leader) in

the jth dimension Fj is the position of the food source in thejth dimension ubj indicates the upper bound of jth di-mension lbj indicates the lower bound of jth dimension andc1 cthe2 and c3 are random numbers

Equation (10) shows that the leader only updates itsposition with respect to the food source +e coefficient c1 isthe most important parameter in SSA because it balancesexploration and exploitation defined as follows [37]

c1 2eminus(4lL)2

(11)

where l is the current iteration and L is the maximumnumber of iterations

+e parameters c2 and c3 are random numbers uni-formly generated in the interval of [0 1] In fact they dictateif the next position in jth dimension should be towardspositive infinity or negative infinity as well as the step size Toupdate the position of the followers the following equationsis utilized (Newtonrsquos law of motion) [37]

xij

12

at2

+ v0t (12)

where ige 2 xij shows the position of ith follower Salp in jth

dimension t is time v0 is the initial speed and a vfinalv0where v x minus x0t

Because the time in optimization is an iteration thediscrepancy between iterations is equal to 1 and consideringv0 0 this equation can be expressed as follows [37]

xij

12

xij + x

iminus1j1113872 1113873 (13)

where ige 2 and xij shows the position of ith follower Salp in

jth dimension Using equations (1) and (4) the Salp chainscan be simulated Figure 2 shows the pseudocode to im-plement the SSA algorithm

4 Formulation of WSN Localization Problem

WSN node localization problem is formulated using thesingle hop range-based distribution technique to estimatethe position of the unknown node coordinates (X Y) withthe aid of anchor nodes (position of known nodes) co-ordinates (x y) Anchor nodes are provided with a GPSdevice so it has the capability of automatically determiningits position Most of the nodes in the WSN are not equippedwith GPS due to high cost To measure the coordinates of Nunknown nodes the procedure followed is given below

Step 1 Randomly initialize the N unknown nodes and Manchor nodes within the communication range (R) Anchornodes measure their position and communicate their co-ordinates to their neighbors For all iterations the nodewhich settles at the end is termed as the reference node andthis node will act as anchor node

Step 2 +ree or more anchor nodes within the commu-nication range of a node are considered as a localized node

Step 3 Let (x y) be the coordinates of the target node to bedetermined and di be the distance between the target nodeand the ith anchor node

di

xminus xi( 11138572

+ yminusyi( 11138572

1113969

(14)

Step 4 +e optimization problem is formulated to minimizethe error of the localization problem +e objective functionfor the localization problem is formulated as

(a) (b)

Figure 1 (a) Individual Salp (b) Salps chain [37]

6 Journal of Computer Networks and Communications

f(x y) min sumM

i1

xminus xi( )2 + yminusyi( )2

radic( )

2 (15)

whereM is the anchor nodes within the transmission range(R) of the target node

Step 5 All target localized nodes (NL) are determined thewhole localization error is calculated as the mean of thesquare of distances of the estimated coordinates node (xi yi)and the actual node coordinates (Xi Yi) for i 1 2 NL

EL 1NL

sumL

i1

xi minusXi( )2 + yi minusYi( )2

radic( ) (16)

e performance of SSA algorithm evaluated using ELand the number of nonlocalized nodes NNL whereNNL [NminusNL]

Step 6 Repeat the steps 2ndash5 until all unknowntarget nodesget localized or no more nodes can be localized

5 Experimental Analysis

In this section the proposed WSN localization approach isevaluated under dierent scenarios and its performance iscompared to four other swarm-based algorithms (PSOBOA FA and GWO) in terms of localization accuracy andcomputing time e computations of the dierent algo-rithms are performed using MATLAB R2012b on a machineof Intel Core i7 CPU 4GB RAM and Windows7 operatingsystem e parametersrsquo values of the deployment area areshown in Table 1

For BOA the sensory modality c is set 001 whereasthe initial value of power exponent a is set to 01 [25] ForPSO initial values of ω 07 and c1 c2 1494 were rec-ommended for faster convergence after experimental tests[25] For FA the randomization parameter α is set to 025the absorption coecient c is set to 10 and the initial

attractiveness parameter β is set to 1 For GWO the pa-rameter a linearly decreases in the interval of [2 to 0] and theC parameter linearly increases from 0 to 2 [6] Finally forSSA c1 is calculated using equation (2) whereas c2 07 andc3 03 and c2 c3 are random numbers uniformly generatedfrom the interval [0 1]

51 Sensor Nodes Localization Using SSA In all conductedexperiments the coordinates of central nodes (N) anddestination nodes (M) are randomly congured during theconstruction of the deployment area e deployment areaincludes three types of nodes anchor nodes whose knownposition target nodes whose unknown position and lo-calized nodes whose positions are already estimated In thissection the performance of the SSA-based localization al-gorithm is evaluated under dierent scenarios using dif-ferent numbers of target nodes and dierent numbers ofanchors as shown in Figure 3

52 Comparison among Dierent Localization AlgorithmsIn this section SSA and the other swarm algorithms havebeen evaluated under dierent scenarios in terms of local-ization error computing time and number of localizednodes e obtained results of the dierent algorithms areshown in Table 2

Under the dierent scenarios (number of nodesnumberof anchors) it is noticed that for all the localization algo-rithms increasing the number of iterations increases both ofthe number of localized nodes and the computing time whilereduces the localization error is notice is rational because

Initialize the Salp population xi (i = 1 2 hellip n) considering ub and lbWhile (end condition is not satisfied)Calculate the fitness of each search agent (salp)F = the best search agentUpdate c1 by equation (2)

Update the position of the leading Salp by equation (1)Else

Update the position of the follower Salp by equation (4)End

End

EndReturn F

Amend the Salps based on the upper and lower bounds of variables

If (i == 1)For each Salp (xi)

Figure 2 Pseudocode of the SSA algorithm

Table 1 Parameters setting of simulation environment

Parameters ValuesSensor nodes Varies on sum6

i1ilowast 25Anchor nodes Varies on increment i i + 5Node transmission range (R) 30mDeployment area 100mtimes 100mMaximum number of iterations 100

Journal of Computer Networks and Communications 7

increasing the number of iterations means higher amountsof computations which requires longer computation timeOn the other side increasing the number of iterations meansthat the chance to nd a better solution get bigger hence thenumber of localized nodes get larger and the value of lo-calization error get smaller For better results analysis theexperimental results are summarized in Table 3

Based on Table 3 regarding the mean localization error(EL) there is no clear pattern that can be detected to rep-resent the relationship between this performance metric andthe number of target nodes and anchor nodes However it isnoticed that SSA has the best results regarding this per-formance metric compared to PSO BOA FA and GWOparticularly when the numbers of target nodes and anchor

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(a)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(b)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(c)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(d)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(e)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(f )

FIGURE 3 Node localization using dierent numbers of target nodes and anchor nodes (a) Target nodes 25 anchor nodes 10 (b) Targetnodes 50 anchor nodes 15 (c) Target nodes 75 anchor nodes 20 (d) Target nodes 100 anchor nodes 25 (e) Target nodes 125anchor nodes 30 (f ) Target nodes 150 anchor nodes 35

8 Journal of Computer Networks and Communications

nodes are increased Regarding the computing time it isnoticed that increasing the number of target nodes andanchor nodes increase the computing time for all local-ization algorithms However once again SSA has the bestcomputing time compared to PSO BOA FA and GWOFinally regarding the number of localized nodes (NL) it isnoticed that SSA has the best results compared to otherlocalization algorithms In addition it is noticed that SSAhas inferior EL when the percentage of (number of anchornodesnumber of target nodes) becomes larger such as(1025) in the first case in Table 3 +at is because thelocalization accuracy increases when the anchor densityincreases with respect to the number of the target nodes[25]

+e graphical representations of the experimentalresults for the different performance metrics are shown inFigures 4ndash6

6 Conclusion

Accurate node localization concerns many applications thatadopt WSNs In this paper a node localization algorithm hasbeen proposed based on a novel bioinspired algorithm calledSalp Swarm Algorithm (SSA) which handled the node lo-calization problem as an optimization problem+e proposedalgorithm has been implemented and validated in differentWSN deployments using different numbers of target nodesand anchor nodes Moreover the proposed algorithm hasbeen evaluated and compared to four well-known optimi-zation algorithms namely PSO BOA FA and GWO interms of localization accuracy computing time and severallocalized nodes +e obtained simulation results have provedthe superiority of the proposed algorithm compared to theother localization algorithms regarding the different perfor-mance metrics In the future work the proposed approach

Table 2 Performance metrics of different localization algorithms

Targetnodes

Anchornodes

No ofiterations

PSO BOA FA GWO SSAEL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10

25 0818 040 16 0232 037 22 0265 06 19 0744 022 20 0465 035 2250 0812 040 15 0225 038 23 0262 12 20 0741 041 21 0462 035 2375 0803 041 18 0223 039 25 0258 15 21 0741 054 21 0458 036 23100 0792 041 18 0221 040 25 0251 18 19 0740 079 23 0451 037 24

50 15

25 0419 071 41 0338 084 47 0477 16 46 0690 042 44 0477 067 4350 0426 073 47 0332 085 48 0473 19 49 0688 063 45 0472 069 4775 0429 076 46 0326 086 49 0465 25 49 0686 081 46 0468 069 48100 0434 076 48 0323 086 49 0465 35 48 0682 098 48 0464 070 50

75 20

25 0735 131 73 0257 149 66 0519 29 71 0641 072 72 0519 090 6950 0728 132 74 0257 149 66 0513 38 72 0641 095 72 0513 092 7275 0728 133 74 0255 150 70 0504 47 73 0638 13 73 0504 095 73100 0724 135 75 0253 152 72 0503 52 73 0635 14 74 0503 096 75

100 25

25 0661 210 97 0355 240 97 0711 38 98 0611 11 95 0511 131 9850 0658 216 97 0355 244 99 0709 42 98 0606 15 97 0509 133 9875 0642 217 99 0333 247 100 0702 56 99 0602 18 98 0502 136 99100 0641 220 100 0331 250 100 0704 63 98 0602 21 98 0504 137 100

125 30

25 0754 487 120 0549 384 122 0829 27 122 0589 15 122 0529 167 12350 0748 486 121 0548 385 123 0824 45 123 0580 22 123 0524 168 12475 0750 489 122 0534 386 124 0822 59 125 0580 28 123 0522 170 125100 0752 495 125 0534 389 124 0822 65 124 0572 33 125 0522 172 125

150 35

25 0625 541 145 0766 588 147 0911 25 149 0559 28 148 0511 212 14950 0622 542 146 0765 561 148 0909 42 150 0547 36 149 0509 214 14975 0619 544 148 0763 564 149 0904 64 150 0523 43 150 0504 216 150100 0616 545 150 0763 569 149 0904 72 149 0523 48 150 0504 218 150

Table 3 Summary of experimental results of the different localization algorithms

Target nodes Anchor nodesPSO BOA FA GWO SSA

EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10 079 040 18 022 038 25 025 18 19 074 054 21 045 035 2450 15 043 076 46 032 086 49 042 25 49 069 081 46 046 069 4875 20 072 135 75 025 152 68 051 47 73 064 095 72 050 096 73100 25 065 216 100 035 245 100 070 53 98 060 21 98 051 135 100125 30 074 490 123 054 387 124 082 65 124 058 28 123 052 170 125150 35 062 543 149 076 565 149 090 72 149 052 43 150 050 215 150

Journal of Computer Networks and Communications 9

0

01

02

03

04M

ean

loca

lizat

ion

erro

r (m

)05

06

07

08

09

1

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

Figure 4 e localization error of the dierent localization algorithms in dierent WSN deployments

0

1

2

3

4

5

6

7

8

9

10

Com

putin

g tim

e (S)

(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35

Figure 5 e computing time of the dierent localization algorithms in dierent WSN deployments

0 20 40 60 80 100 120 140Number of localized nodes

(Num

ber o

f tar

get n

odes

ndashnu

mbe

r of

anch

ors)

SSAGWOFA

BOAPSO

25 ndash 10

50 ndash 15

75 ndash 20

100 ndash 25

125 ndash 30

150 ndash 35

Figure 6 e number of localized nodes of the dierent localization algorithms in dierent WSN deployments

10 Journal of Computer Networks and Communications

can be hybridized with other algorithm to reduce the local-ization error

Data Availability

No data were used to support this study Because our articlediscusses the problem of localization in WSNs our exper-imental results have been applied based on several networksrsquosize

Conflicts of Interest

+e authors declare that they have no conflicts of interest

References

[1] R V Kulkarni G K Venayagamoorthy and M X ChengldquoBio-inspired node localization in wireless sensor networksrdquoin Proceedings of IEEE International Conference in SystemsMan and Cybernetics (SMC) pp 205ndash210 San Antonio TXUSA October 2009

[2] D Lavanya and S K Udgata ldquoSwarm intelligence-basedlocalization in wireless sensor networksrdquo in InternationalWorkshop on Multi-Disciplinary Trends in Artificial In-telligence pp 317ndash328 Springer Berlin Germany 2011

[3] J Wang R K Ghosh and S K Das ldquoA survey on sensorlocalizationrdquo Journal of Control Aeory and Applicationsvol 8 no 1 pp 2ndash11 2010

[4] J Aspnes T Eren D K Goldenberg et al ldquoA theory ofnetwork localizationrdquo IEEE Transactions on Mobile Com-puting vol 5 no 12 pp 1663ndash1678 2006

[5] S Goyal and M S Patterh ldquoModified bat algorithm for lo-calization of wireless sensor networkrdquo Wireless PersonalCommunications vol 86 no 2 pp 657ndash670 2015

[6] M M Ahmed E H Houssein A E Hassanien A Taha andE Hassanien ldquoMaximizing lifetime of wireless sensor net-works based on whale optimization algorithmrdquo in In-ternational Conference on Advanced Intelligent Systems andInformatics pp 724ndash733 Springer Berlin Germany 2017

[7] K S Low H A Nguyen and H Guo ldquoA particle swarmoptimization approach for the localization of a wireless sensornetworkrdquo in Proceedings of IEEE International Symposium inIndustrial Electronics (ISIE) pp 1820ndash1825 VancouverCanada June 2008

[8] D Manjarres J Del Ser S Gil-Lopez M Vecchio I Landa-Torres and R Lopez-Valcarce ldquoOn the application of a hybridharmony search algorithm to node localization in anchor-based wireless sensor networksrdquo in Proceedings of IEEE In-ternational Conference in Intelligent Systems Design andApplications (ISDA) pp 1014ndash1019 Cordoba Spain No-vember 2011

[9] M Li W Xiong and Q Liang ldquoAn improved abc-based nodelocalization algorithm for wireless sensor networkrdquo in Pro-ceedings of IEEE International Conference in Wireless Com-munications Networking and Mobile Computing (WiCOM)pp 1ndash4 Barcelona Spain September 2012

[10] A Tamizharasi R Arthi and K Murugan ldquoBio-inspiredalgorithm for optimizing the localization of wireless sensornetworksrdquo in Proceedings of IEEE International Conference inComputing Communications and Networking Technologies(ICCCNT) pp 1ndash5 Tiruchengode India June 2013

[11] C Tang R Liu and J Ni ldquoA novel wireless sensor networklocalization approach localization based on plant growth

simulation algorithmrdquo Elektronika ir Elektrotechnika vol 19no 8 pp 97ndash100 2013

[12] O D Jegede and K Ferens ldquoA genetic algorithm for nodelocalization in wireless sensor networksrdquo in Proceedings ofInternational Conference on Genetic and EvolutionaryMethods (GEM) Ae Steering Committee of the World Con-gress in Computer Science Computer Engineering and AppliedComputing (WorldComp) pp 1ndash7 Las Vegas NV USA July2013

[13] S Goyal and M S Patterh ldquoWireless sensor network local-ization based on cuckoo search algorithmrdquo Wireless personalcommunications vol 79 no 1 pp 223ndash234 2014

[14] L Dan and W Xian-bin ldquoAn improved PSO algorithm fordistributed localization in wireless sensor networksrdquo In-ternational Journal of Distributed Sensor Networks vol 11no 7 Article ID 970272 2015

[15] R Krishnaprabha and A Gopakumar ldquoPerformance ofgravitational search algorithm in wireless sensor networklocalizationrdquo in Proceedings of IEEE National Conference inCommunication Signal Processing and Networking (NCCSN)pp 1ndash6 Palakkad India October 2014

[16] B Peng and L Li ldquoAn improved localization algorithm basedon genetic algorithm in wireless sensor networksrdquo CognitiveNeurodynamics vol 9 no 2 pp 249ndash256 2015

[17] V O Sai C S Shieh T T Nguyen Y C Lin M F Horngand Q D Le ldquoParallel firefly algorithm for localization al-gorithm in wireless sensor networkrdquo in Proceedings of IEEEInternational Conference in Robot Vision and Signal Pro-cessing (RVSP) pp 300ndash305 Kaohsiung China November2015

[18] S Sivakumar and R Venkatesan ldquoMeta-heuristic approachesfor minimizing error in localization of wireless sensor net-worksrdquo Applied Soft Computing vol 36 pp 506ndash518 2015

[19] Z Sun L Tao X Wang and Z Zhou ldquoLocalization algorithmin wireless sensor networks based on multiobjective particleswarm optimizationrdquo International Journal of DistributedSensor Networks vol 11 no 8 Article ID 716291 2015

[20] A Arsic M Tuba and M Jordanski ldquoFireworks algorithmapplied to wireless sensor networks localization problemrdquo inProceedings of IEEE Congress on Evolutionary Computation(CEC) pp 4038ndash4044 Vancouver Canada July 2016

[21] C S Shieh V O Sai Y C Lin T F Lee T T Nguyen andQ D Le ldquoImproved node localization for WSN using heu-ristic optimization approachesrdquo in Proceedings of IEEE In-ternational Conference in Networking and NetworkApplications (NaNA) pp 95ndash98 Hokkaido Japan July 2016

[22] J Cheng and L Xia ldquoAn effective Cuckoo search algorithm fornode localization in wireless sensor networkrdquo Sensors vol 16no 9 p 1390 2016

[23] P T Daely and S Y Shin ldquoRange based wireless node lo-calization using Dragonfly Algorithmrdquo in Proceedings of IEEEInternational Conference in Ubiquitous and Future Networks(ICUFN) pp 1012ndash1015 Vienna Austria July 2016

[24] T T Nguyen J S Pan S C Chu J F Roddick and T K DaoldquoOptimization localization in wireless sensor network basedon multi-objective firefly algorithmrdquo Journal of NetworkIntelligence vol 1 no 4 pp 130ndash138 2016

[25] S Arora and S Singh ldquoNode localization in wireless sensornetworks using butterfly optimization algorithmrdquo ArabianJournal for Science and Engineering vol 42 no 8pp 3325ndash3335 2017

[26] S R Sujatha and M Siddappa ldquoNode localization method forwireless sensor networks based on hybrid optimization ofparticle swarm optimization and differential evolutionrdquo IOSR

Journal of Computer Networks and Communications 11

Journal of Computer Engineering vol 19 no 2 pp 07ndash122017

[27] R Kaur and S Arora ldquonature inspired range based wirelesssensor node localization algorithmsrdquo International Journal ofInteractive Multimedia and Artificial Intelligence vol 4pp 7ndash17 2017

[28] A +arwat E H Houssein M M Ahmed A E Hassanienand T Gabel ldquoMOGOA algorithm for constrained and un-constrained multi-objective optimization problemsrdquo AppliedIntelligence vol 48 no 8 pp 2268ndash2283 2017

[29] A A Ewees M Abd Elaziz and E H Houssein ldquoImprovedgrasshopper optimization algorithm using opposition-basedlearningrdquo Expert Systems with Applications vol 112pp 156ndash172 2018

[30] A G Hussien E H Houssein and A E Hassanien ldquoA binarywhale optimization algorithm with hyperbolic tangent fitnessfunction for feature selectionrdquo in Proceedings of Eighth In-ternational Conference on Intelligent Computing and In-formation Systems (ICICIS) pp 166ndash172 Cairo EgyptDecember 2017

[31] B Xue M Zhang and W N Browne ldquoParticle swarm op-timization for feature selection in classification a multi-objective approachrdquo IEEE Transactions on Cyberneticsvol 43 no 6 pp 1656ndash1671 2013

[32] S Arora and S Singh ldquoButterfly algorithmwith levy flights forglobal optimizationrdquo in Proceedings of IEEE InternationalConference in Signal Processing Computing and Control(ISPCC) pp 220ndash224 Waknaghat India September 2015

[33] S Arora and S Singh ldquoAn improved butterfly optimizationalgorithm with chaosrdquo Journal of Intelligent and Fuzzy Sys-tems vol 32 no 1 pp 1079ndash1088 2017

[34] R B Blair and A E Launer ldquoButterfly diversity and humanland use species assemblages along an urban grandientrdquoBiological Conservation vol 80 no 1 pp 113ndash125 1997

[35] X S Yang and X He ldquoFirefly algorithm recent advances andapplicationsrdquo 2013 httpsarxivorgabs13083898

[36] X S Yang ldquoFirefly algorithms for multimodal optimizationrdquoin Proceedings of International Symposium on Stochastic Al-gorithms pp 169ndash178 Sapporo Japan October 2009

[37] S Mirjalili S M Mirjalili and A Lewis ldquoGrey wolf opti-mizerrdquo Advances in Engineering Software vol 69 pp 46ndash612014

[38] L P Madin ldquoAspects of jet propulsion in Salpsrdquo CanadianJournal of Zoology vol 68 no 4 pp 765ndash777 1990

[39] S Mirjalili A H Gandomi S Z Mirjalili S Saremi H Farisand S M Mirjalili ldquoSalp swarm algorithm a bio-inspiredoptimizer for engineering design problemsrdquo Advances inEngineering Software vol 114 pp 163ndash191 2017

[40] P A Anderson and Q Bone ldquoCommunication betweenindividuals in Salp chains II Physiologyrdquo Proceedings of theRoyal Society of London Series B Biological Sciences vol 210no 1181 pp 559ndash574 1980

[41] H T Ibrahim W J Mazher O N Ucan and O BayatldquoFeature selection using salp swarm algorithm for real bio-medical datasetsrdquo IJCSNS vol 17 no 12 pp 13ndash20 2017

12 Journal of Computer Networks and Communications

fi is the fragrance of ith butterfly and r is a random numberin [0 1] Local search phase is formulated as

xt+1i x

ti + r

2times x

tk minusx

tj1113872 1113873 times fi (5)

where xtj and xt

k are jth and kth butterflies chosen randomlyfrom the solution space Equation (5) is considered a localrandom walk if and only if xt

j and xtk belongs to the same

subswarm and r is a random number in [0 1] Search forfood and mating partner by butterflies can occur at bothlocal and global scale therefore a switch probability p isemployed in BOA to switch between common global searchand intensive local search

33 Firefly Algorithm Firefly algorithm (FA) is a nature-inspired algorithm proposed by Yang and He [35] It mimicsthe social behaviors and flashing patterns of fireflies FAdepends on the next three idealized rules [36]

(i) Fireflies are unisex whichmeans that a firefly can getattracted to any other firefly regardless their sex

(ii) +e attractiveness of fireflies is directly proportionalto their brightness +us for any two flashingfireflies the firefly with less brightness moves to-ward the one with higher brightness If there is no abrighter firefly than a particular firefly the lattermoves randomly

(iii) +e brightness of a firefly is calculated using anobjective function

A fireflyrsquos attractiveness is proportional to the light in-tensity visualized by other fireflies in the region thereforethe relationship between the attractiveness β and the dis-tance r can be formulated as

β β0eminuscr2

(6)

where β0 is the brightness at distance r 0 and c is the lightabsorption coefficient +e movement of a firefly i toward amore attractive (brighter) firefly j is represented as

xt+1i x

ti + β0e

minuscr2ij x

tj minus x

ti1113872 1113873 + αtϵ

ti (7)

where rij is the distance between the fireflies i and j that iscalculated using the Euclidean norm the second term is dueto the attraction the third term is a randomization with αt

being the randomization parameter and ϵti is a vector thatincludes random numbers

34 Grey Wolf Optimizer Grey wolf optimizer (GWO) is arecent swarm intelligence algorithm inspired by the greywolf community It is developed by Mirjalili et al in 2014Grey wolf is a very dangerous creature which belongs to theCanidae family Grey wolves usually live in packs that consistof 5 to 12 wolves Each group has social dominance hier-archy alpha beta and omega in order +e alphas are amale and female which represent the leaders of the group+e betas are the second level of management hierarchy+eomegas are the final level in the hierarchy [37]

In order to mathematically model the social hierarchy ofwolves in GWO the fittest solution is referred to as the alpha(α) Consequently the second and third best solutions arebeta (szlig) and delta (δ) respectively +e rest of candidatesolutions are omega (ω) [35]+emathematical model of theencircling behavior is represented as follows [35]

Drarr

| Crarr

Xrarr

p (t)minus Xrarr

(t)|

Xrarr

(t + 1) |Xrarr

p(t)minus Ararr

Drarr

|(8)

where t indicates the current iteration Crarr

2 rrarr

2Ararr

2 ararr

rrarr

1 minus ararr X

rarrm is the position vector of the wolf r1

and r2 are random vectors in [0 1] and linearly varies from 2to 1 C and A are coefficient vectors and Xp is the positionvector of the prey During the optimization ωwolves updatetheir positions around αβ and δ +erefore the ω wolvescan reposition with respect to α β and δ as shown below[35]

Drarrα C

rarr1X

rarrαminus X

rarr11138681113868111386811138681113868

11138681113868111386811138681113868Drarrβ C

rarr2X

rarrβminus X

rarr11138681113868111386811138681113868

11138681113868111386811138681113868Drarrδ

Crarr

3Xrarrδ minus X

rarr11138681113868111386811138681113868

11138681113868111386811138681113868

Xrarr1 X

rarrαminus A

rarr1(D

rarrα)X

rarr2 X

rarrβminus A

rarr2(D

rarrβ)X

rarr3

Xrarrδ minus A

rarr3(D

rarrδ)

Xrarr

(t + 1) Xrarr

1 + Xrarr

2 + Xrarr

3

3

(9)

With these equations a search agent updates its positionaccording to α β and δ in an n-dimensional search space Inaddition the final position would be in a random placewithin a circle which is defined by the positions of αβ and δin the search space [37]

35 Salp SwarmAlgorithm Salps are part of Salpidae familywith the limpid cylinder design body +ey look like jelly-fishes in texture andmovement+e shape of a Salp is shownin Figure 1(a) +e water is pushed Salps bodies to moveforward [38] Generally the biological research about Salpsis still in its early stages because their living environmentsare hardly accessible and it is very difficult to keep them inlaboratory environment Salps swarming attitude is the maininspiration to build Salp swarm algorithm [39] Salpscompose a swarm in profound oceans which is called Salpchain +is chain is illustrated in Figure 1(b) +is chain canhelp to achieve better locomotion during the foragingprocess [40]

Originally the Salps population is divided into twogroups to formulate the mathematical model for Salp chainshead and followers +e head position is at the beginning ofthe chain while the rest of the chain is referred to as thefollowers [41]

+e Salps location is determined likewise swarm-basedmethods by an n-dimensional search area via considering

Journal of Computer Networks and Communications 5

the number of variables inside the presented problemrepresented by n Accordingly a two-dimensional matrixdescribed as x will reservation the position of all Salps It issupposed too in search space there is ldquoFrdquo which is a foodsource as the target of the swarm +e following equation issuggested to upgrade the leader location [37]

x1j

Fj + c1 ubj minus lbj1113872 1113873c2 + lbj1113872 1113873 c3 ge 0

Fj minus c1 ubj minus lbj1113872 1113873c2 + lbj1113872 1113873 c3 lt 0

⎧⎪⎨

⎪⎩(10)

where x1j shows the position of the first Salp (leader) in

the jth dimension Fj is the position of the food source in thejth dimension ubj indicates the upper bound of jth di-mension lbj indicates the lower bound of jth dimension andc1 cthe2 and c3 are random numbers

Equation (10) shows that the leader only updates itsposition with respect to the food source +e coefficient c1 isthe most important parameter in SSA because it balancesexploration and exploitation defined as follows [37]

c1 2eminus(4lL)2

(11)

where l is the current iteration and L is the maximumnumber of iterations

+e parameters c2 and c3 are random numbers uni-formly generated in the interval of [0 1] In fact they dictateif the next position in jth dimension should be towardspositive infinity or negative infinity as well as the step size Toupdate the position of the followers the following equationsis utilized (Newtonrsquos law of motion) [37]

xij

12

at2

+ v0t (12)

where ige 2 xij shows the position of ith follower Salp in jth

dimension t is time v0 is the initial speed and a vfinalv0where v x minus x0t

Because the time in optimization is an iteration thediscrepancy between iterations is equal to 1 and consideringv0 0 this equation can be expressed as follows [37]

xij

12

xij + x

iminus1j1113872 1113873 (13)

where ige 2 and xij shows the position of ith follower Salp in

jth dimension Using equations (1) and (4) the Salp chainscan be simulated Figure 2 shows the pseudocode to im-plement the SSA algorithm

4 Formulation of WSN Localization Problem

WSN node localization problem is formulated using thesingle hop range-based distribution technique to estimatethe position of the unknown node coordinates (X Y) withthe aid of anchor nodes (position of known nodes) co-ordinates (x y) Anchor nodes are provided with a GPSdevice so it has the capability of automatically determiningits position Most of the nodes in the WSN are not equippedwith GPS due to high cost To measure the coordinates of Nunknown nodes the procedure followed is given below

Step 1 Randomly initialize the N unknown nodes and Manchor nodes within the communication range (R) Anchornodes measure their position and communicate their co-ordinates to their neighbors For all iterations the nodewhich settles at the end is termed as the reference node andthis node will act as anchor node

Step 2 +ree or more anchor nodes within the commu-nication range of a node are considered as a localized node

Step 3 Let (x y) be the coordinates of the target node to bedetermined and di be the distance between the target nodeand the ith anchor node

di

xminus xi( 11138572

+ yminusyi( 11138572

1113969

(14)

Step 4 +e optimization problem is formulated to minimizethe error of the localization problem +e objective functionfor the localization problem is formulated as

(a) (b)

Figure 1 (a) Individual Salp (b) Salps chain [37]

6 Journal of Computer Networks and Communications

f(x y) min sumM

i1

xminus xi( )2 + yminusyi( )2

radic( )

2 (15)

whereM is the anchor nodes within the transmission range(R) of the target node

Step 5 All target localized nodes (NL) are determined thewhole localization error is calculated as the mean of thesquare of distances of the estimated coordinates node (xi yi)and the actual node coordinates (Xi Yi) for i 1 2 NL

EL 1NL

sumL

i1

xi minusXi( )2 + yi minusYi( )2

radic( ) (16)

e performance of SSA algorithm evaluated using ELand the number of nonlocalized nodes NNL whereNNL [NminusNL]

Step 6 Repeat the steps 2ndash5 until all unknowntarget nodesget localized or no more nodes can be localized

5 Experimental Analysis

In this section the proposed WSN localization approach isevaluated under dierent scenarios and its performance iscompared to four other swarm-based algorithms (PSOBOA FA and GWO) in terms of localization accuracy andcomputing time e computations of the dierent algo-rithms are performed using MATLAB R2012b on a machineof Intel Core i7 CPU 4GB RAM and Windows7 operatingsystem e parametersrsquo values of the deployment area areshown in Table 1

For BOA the sensory modality c is set 001 whereasthe initial value of power exponent a is set to 01 [25] ForPSO initial values of ω 07 and c1 c2 1494 were rec-ommended for faster convergence after experimental tests[25] For FA the randomization parameter α is set to 025the absorption coecient c is set to 10 and the initial

attractiveness parameter β is set to 1 For GWO the pa-rameter a linearly decreases in the interval of [2 to 0] and theC parameter linearly increases from 0 to 2 [6] Finally forSSA c1 is calculated using equation (2) whereas c2 07 andc3 03 and c2 c3 are random numbers uniformly generatedfrom the interval [0 1]

51 Sensor Nodes Localization Using SSA In all conductedexperiments the coordinates of central nodes (N) anddestination nodes (M) are randomly congured during theconstruction of the deployment area e deployment areaincludes three types of nodes anchor nodes whose knownposition target nodes whose unknown position and lo-calized nodes whose positions are already estimated In thissection the performance of the SSA-based localization al-gorithm is evaluated under dierent scenarios using dif-ferent numbers of target nodes and dierent numbers ofanchors as shown in Figure 3

52 Comparison among Dierent Localization AlgorithmsIn this section SSA and the other swarm algorithms havebeen evaluated under dierent scenarios in terms of local-ization error computing time and number of localizednodes e obtained results of the dierent algorithms areshown in Table 2

Under the dierent scenarios (number of nodesnumberof anchors) it is noticed that for all the localization algo-rithms increasing the number of iterations increases both ofthe number of localized nodes and the computing time whilereduces the localization error is notice is rational because

Initialize the Salp population xi (i = 1 2 hellip n) considering ub and lbWhile (end condition is not satisfied)Calculate the fitness of each search agent (salp)F = the best search agentUpdate c1 by equation (2)

Update the position of the leading Salp by equation (1)Else

Update the position of the follower Salp by equation (4)End

End

EndReturn F

Amend the Salps based on the upper and lower bounds of variables

If (i == 1)For each Salp (xi)

Figure 2 Pseudocode of the SSA algorithm

Table 1 Parameters setting of simulation environment

Parameters ValuesSensor nodes Varies on sum6

i1ilowast 25Anchor nodes Varies on increment i i + 5Node transmission range (R) 30mDeployment area 100mtimes 100mMaximum number of iterations 100

Journal of Computer Networks and Communications 7

increasing the number of iterations means higher amountsof computations which requires longer computation timeOn the other side increasing the number of iterations meansthat the chance to nd a better solution get bigger hence thenumber of localized nodes get larger and the value of lo-calization error get smaller For better results analysis theexperimental results are summarized in Table 3

Based on Table 3 regarding the mean localization error(EL) there is no clear pattern that can be detected to rep-resent the relationship between this performance metric andthe number of target nodes and anchor nodes However it isnoticed that SSA has the best results regarding this per-formance metric compared to PSO BOA FA and GWOparticularly when the numbers of target nodes and anchor

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(a)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(b)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(c)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(d)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(e)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(f )

FIGURE 3 Node localization using dierent numbers of target nodes and anchor nodes (a) Target nodes 25 anchor nodes 10 (b) Targetnodes 50 anchor nodes 15 (c) Target nodes 75 anchor nodes 20 (d) Target nodes 100 anchor nodes 25 (e) Target nodes 125anchor nodes 30 (f ) Target nodes 150 anchor nodes 35

8 Journal of Computer Networks and Communications

nodes are increased Regarding the computing time it isnoticed that increasing the number of target nodes andanchor nodes increase the computing time for all local-ization algorithms However once again SSA has the bestcomputing time compared to PSO BOA FA and GWOFinally regarding the number of localized nodes (NL) it isnoticed that SSA has the best results compared to otherlocalization algorithms In addition it is noticed that SSAhas inferior EL when the percentage of (number of anchornodesnumber of target nodes) becomes larger such as(1025) in the first case in Table 3 +at is because thelocalization accuracy increases when the anchor densityincreases with respect to the number of the target nodes[25]

+e graphical representations of the experimentalresults for the different performance metrics are shown inFigures 4ndash6

6 Conclusion

Accurate node localization concerns many applications thatadopt WSNs In this paper a node localization algorithm hasbeen proposed based on a novel bioinspired algorithm calledSalp Swarm Algorithm (SSA) which handled the node lo-calization problem as an optimization problem+e proposedalgorithm has been implemented and validated in differentWSN deployments using different numbers of target nodesand anchor nodes Moreover the proposed algorithm hasbeen evaluated and compared to four well-known optimi-zation algorithms namely PSO BOA FA and GWO interms of localization accuracy computing time and severallocalized nodes +e obtained simulation results have provedthe superiority of the proposed algorithm compared to theother localization algorithms regarding the different perfor-mance metrics In the future work the proposed approach

Table 2 Performance metrics of different localization algorithms

Targetnodes

Anchornodes

No ofiterations

PSO BOA FA GWO SSAEL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10

25 0818 040 16 0232 037 22 0265 06 19 0744 022 20 0465 035 2250 0812 040 15 0225 038 23 0262 12 20 0741 041 21 0462 035 2375 0803 041 18 0223 039 25 0258 15 21 0741 054 21 0458 036 23100 0792 041 18 0221 040 25 0251 18 19 0740 079 23 0451 037 24

50 15

25 0419 071 41 0338 084 47 0477 16 46 0690 042 44 0477 067 4350 0426 073 47 0332 085 48 0473 19 49 0688 063 45 0472 069 4775 0429 076 46 0326 086 49 0465 25 49 0686 081 46 0468 069 48100 0434 076 48 0323 086 49 0465 35 48 0682 098 48 0464 070 50

75 20

25 0735 131 73 0257 149 66 0519 29 71 0641 072 72 0519 090 6950 0728 132 74 0257 149 66 0513 38 72 0641 095 72 0513 092 7275 0728 133 74 0255 150 70 0504 47 73 0638 13 73 0504 095 73100 0724 135 75 0253 152 72 0503 52 73 0635 14 74 0503 096 75

100 25

25 0661 210 97 0355 240 97 0711 38 98 0611 11 95 0511 131 9850 0658 216 97 0355 244 99 0709 42 98 0606 15 97 0509 133 9875 0642 217 99 0333 247 100 0702 56 99 0602 18 98 0502 136 99100 0641 220 100 0331 250 100 0704 63 98 0602 21 98 0504 137 100

125 30

25 0754 487 120 0549 384 122 0829 27 122 0589 15 122 0529 167 12350 0748 486 121 0548 385 123 0824 45 123 0580 22 123 0524 168 12475 0750 489 122 0534 386 124 0822 59 125 0580 28 123 0522 170 125100 0752 495 125 0534 389 124 0822 65 124 0572 33 125 0522 172 125

150 35

25 0625 541 145 0766 588 147 0911 25 149 0559 28 148 0511 212 14950 0622 542 146 0765 561 148 0909 42 150 0547 36 149 0509 214 14975 0619 544 148 0763 564 149 0904 64 150 0523 43 150 0504 216 150100 0616 545 150 0763 569 149 0904 72 149 0523 48 150 0504 218 150

Table 3 Summary of experimental results of the different localization algorithms

Target nodes Anchor nodesPSO BOA FA GWO SSA

EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10 079 040 18 022 038 25 025 18 19 074 054 21 045 035 2450 15 043 076 46 032 086 49 042 25 49 069 081 46 046 069 4875 20 072 135 75 025 152 68 051 47 73 064 095 72 050 096 73100 25 065 216 100 035 245 100 070 53 98 060 21 98 051 135 100125 30 074 490 123 054 387 124 082 65 124 058 28 123 052 170 125150 35 062 543 149 076 565 149 090 72 149 052 43 150 050 215 150

Journal of Computer Networks and Communications 9

0

01

02

03

04M

ean

loca

lizat

ion

erro

r (m

)05

06

07

08

09

1

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

Figure 4 e localization error of the dierent localization algorithms in dierent WSN deployments

0

1

2

3

4

5

6

7

8

9

10

Com

putin

g tim

e (S)

(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35

Figure 5 e computing time of the dierent localization algorithms in dierent WSN deployments

0 20 40 60 80 100 120 140Number of localized nodes

(Num

ber o

f tar

get n

odes

ndashnu

mbe

r of

anch

ors)

SSAGWOFA

BOAPSO

25 ndash 10

50 ndash 15

75 ndash 20

100 ndash 25

125 ndash 30

150 ndash 35

Figure 6 e number of localized nodes of the dierent localization algorithms in dierent WSN deployments

10 Journal of Computer Networks and Communications

can be hybridized with other algorithm to reduce the local-ization error

Data Availability

No data were used to support this study Because our articlediscusses the problem of localization in WSNs our exper-imental results have been applied based on several networksrsquosize

Conflicts of Interest

+e authors declare that they have no conflicts of interest

References

[1] R V Kulkarni G K Venayagamoorthy and M X ChengldquoBio-inspired node localization in wireless sensor networksrdquoin Proceedings of IEEE International Conference in SystemsMan and Cybernetics (SMC) pp 205ndash210 San Antonio TXUSA October 2009

[2] D Lavanya and S K Udgata ldquoSwarm intelligence-basedlocalization in wireless sensor networksrdquo in InternationalWorkshop on Multi-Disciplinary Trends in Artificial In-telligence pp 317ndash328 Springer Berlin Germany 2011

[3] J Wang R K Ghosh and S K Das ldquoA survey on sensorlocalizationrdquo Journal of Control Aeory and Applicationsvol 8 no 1 pp 2ndash11 2010

[4] J Aspnes T Eren D K Goldenberg et al ldquoA theory ofnetwork localizationrdquo IEEE Transactions on Mobile Com-puting vol 5 no 12 pp 1663ndash1678 2006

[5] S Goyal and M S Patterh ldquoModified bat algorithm for lo-calization of wireless sensor networkrdquo Wireless PersonalCommunications vol 86 no 2 pp 657ndash670 2015

[6] M M Ahmed E H Houssein A E Hassanien A Taha andE Hassanien ldquoMaximizing lifetime of wireless sensor net-works based on whale optimization algorithmrdquo in In-ternational Conference on Advanced Intelligent Systems andInformatics pp 724ndash733 Springer Berlin Germany 2017

[7] K S Low H A Nguyen and H Guo ldquoA particle swarmoptimization approach for the localization of a wireless sensornetworkrdquo in Proceedings of IEEE International Symposium inIndustrial Electronics (ISIE) pp 1820ndash1825 VancouverCanada June 2008

[8] D Manjarres J Del Ser S Gil-Lopez M Vecchio I Landa-Torres and R Lopez-Valcarce ldquoOn the application of a hybridharmony search algorithm to node localization in anchor-based wireless sensor networksrdquo in Proceedings of IEEE In-ternational Conference in Intelligent Systems Design andApplications (ISDA) pp 1014ndash1019 Cordoba Spain No-vember 2011

[9] M Li W Xiong and Q Liang ldquoAn improved abc-based nodelocalization algorithm for wireless sensor networkrdquo in Pro-ceedings of IEEE International Conference in Wireless Com-munications Networking and Mobile Computing (WiCOM)pp 1ndash4 Barcelona Spain September 2012

[10] A Tamizharasi R Arthi and K Murugan ldquoBio-inspiredalgorithm for optimizing the localization of wireless sensornetworksrdquo in Proceedings of IEEE International Conference inComputing Communications and Networking Technologies(ICCCNT) pp 1ndash5 Tiruchengode India June 2013

[11] C Tang R Liu and J Ni ldquoA novel wireless sensor networklocalization approach localization based on plant growth

simulation algorithmrdquo Elektronika ir Elektrotechnika vol 19no 8 pp 97ndash100 2013

[12] O D Jegede and K Ferens ldquoA genetic algorithm for nodelocalization in wireless sensor networksrdquo in Proceedings ofInternational Conference on Genetic and EvolutionaryMethods (GEM) Ae Steering Committee of the World Con-gress in Computer Science Computer Engineering and AppliedComputing (WorldComp) pp 1ndash7 Las Vegas NV USA July2013

[13] S Goyal and M S Patterh ldquoWireless sensor network local-ization based on cuckoo search algorithmrdquo Wireless personalcommunications vol 79 no 1 pp 223ndash234 2014

[14] L Dan and W Xian-bin ldquoAn improved PSO algorithm fordistributed localization in wireless sensor networksrdquo In-ternational Journal of Distributed Sensor Networks vol 11no 7 Article ID 970272 2015

[15] R Krishnaprabha and A Gopakumar ldquoPerformance ofgravitational search algorithm in wireless sensor networklocalizationrdquo in Proceedings of IEEE National Conference inCommunication Signal Processing and Networking (NCCSN)pp 1ndash6 Palakkad India October 2014

[16] B Peng and L Li ldquoAn improved localization algorithm basedon genetic algorithm in wireless sensor networksrdquo CognitiveNeurodynamics vol 9 no 2 pp 249ndash256 2015

[17] V O Sai C S Shieh T T Nguyen Y C Lin M F Horngand Q D Le ldquoParallel firefly algorithm for localization al-gorithm in wireless sensor networkrdquo in Proceedings of IEEEInternational Conference in Robot Vision and Signal Pro-cessing (RVSP) pp 300ndash305 Kaohsiung China November2015

[18] S Sivakumar and R Venkatesan ldquoMeta-heuristic approachesfor minimizing error in localization of wireless sensor net-worksrdquo Applied Soft Computing vol 36 pp 506ndash518 2015

[19] Z Sun L Tao X Wang and Z Zhou ldquoLocalization algorithmin wireless sensor networks based on multiobjective particleswarm optimizationrdquo International Journal of DistributedSensor Networks vol 11 no 8 Article ID 716291 2015

[20] A Arsic M Tuba and M Jordanski ldquoFireworks algorithmapplied to wireless sensor networks localization problemrdquo inProceedings of IEEE Congress on Evolutionary Computation(CEC) pp 4038ndash4044 Vancouver Canada July 2016

[21] C S Shieh V O Sai Y C Lin T F Lee T T Nguyen andQ D Le ldquoImproved node localization for WSN using heu-ristic optimization approachesrdquo in Proceedings of IEEE In-ternational Conference in Networking and NetworkApplications (NaNA) pp 95ndash98 Hokkaido Japan July 2016

[22] J Cheng and L Xia ldquoAn effective Cuckoo search algorithm fornode localization in wireless sensor networkrdquo Sensors vol 16no 9 p 1390 2016

[23] P T Daely and S Y Shin ldquoRange based wireless node lo-calization using Dragonfly Algorithmrdquo in Proceedings of IEEEInternational Conference in Ubiquitous and Future Networks(ICUFN) pp 1012ndash1015 Vienna Austria July 2016

[24] T T Nguyen J S Pan S C Chu J F Roddick and T K DaoldquoOptimization localization in wireless sensor network basedon multi-objective firefly algorithmrdquo Journal of NetworkIntelligence vol 1 no 4 pp 130ndash138 2016

[25] S Arora and S Singh ldquoNode localization in wireless sensornetworks using butterfly optimization algorithmrdquo ArabianJournal for Science and Engineering vol 42 no 8pp 3325ndash3335 2017

[26] S R Sujatha and M Siddappa ldquoNode localization method forwireless sensor networks based on hybrid optimization ofparticle swarm optimization and differential evolutionrdquo IOSR

Journal of Computer Networks and Communications 11

Journal of Computer Engineering vol 19 no 2 pp 07ndash122017

[27] R Kaur and S Arora ldquonature inspired range based wirelesssensor node localization algorithmsrdquo International Journal ofInteractive Multimedia and Artificial Intelligence vol 4pp 7ndash17 2017

[28] A +arwat E H Houssein M M Ahmed A E Hassanienand T Gabel ldquoMOGOA algorithm for constrained and un-constrained multi-objective optimization problemsrdquo AppliedIntelligence vol 48 no 8 pp 2268ndash2283 2017

[29] A A Ewees M Abd Elaziz and E H Houssein ldquoImprovedgrasshopper optimization algorithm using opposition-basedlearningrdquo Expert Systems with Applications vol 112pp 156ndash172 2018

[30] A G Hussien E H Houssein and A E Hassanien ldquoA binarywhale optimization algorithm with hyperbolic tangent fitnessfunction for feature selectionrdquo in Proceedings of Eighth In-ternational Conference on Intelligent Computing and In-formation Systems (ICICIS) pp 166ndash172 Cairo EgyptDecember 2017

[31] B Xue M Zhang and W N Browne ldquoParticle swarm op-timization for feature selection in classification a multi-objective approachrdquo IEEE Transactions on Cyberneticsvol 43 no 6 pp 1656ndash1671 2013

[32] S Arora and S Singh ldquoButterfly algorithmwith levy flights forglobal optimizationrdquo in Proceedings of IEEE InternationalConference in Signal Processing Computing and Control(ISPCC) pp 220ndash224 Waknaghat India September 2015

[33] S Arora and S Singh ldquoAn improved butterfly optimizationalgorithm with chaosrdquo Journal of Intelligent and Fuzzy Sys-tems vol 32 no 1 pp 1079ndash1088 2017

[34] R B Blair and A E Launer ldquoButterfly diversity and humanland use species assemblages along an urban grandientrdquoBiological Conservation vol 80 no 1 pp 113ndash125 1997

[35] X S Yang and X He ldquoFirefly algorithm recent advances andapplicationsrdquo 2013 httpsarxivorgabs13083898

[36] X S Yang ldquoFirefly algorithms for multimodal optimizationrdquoin Proceedings of International Symposium on Stochastic Al-gorithms pp 169ndash178 Sapporo Japan October 2009

[37] S Mirjalili S M Mirjalili and A Lewis ldquoGrey wolf opti-mizerrdquo Advances in Engineering Software vol 69 pp 46ndash612014

[38] L P Madin ldquoAspects of jet propulsion in Salpsrdquo CanadianJournal of Zoology vol 68 no 4 pp 765ndash777 1990

[39] S Mirjalili A H Gandomi S Z Mirjalili S Saremi H Farisand S M Mirjalili ldquoSalp swarm algorithm a bio-inspiredoptimizer for engineering design problemsrdquo Advances inEngineering Software vol 114 pp 163ndash191 2017

[40] P A Anderson and Q Bone ldquoCommunication betweenindividuals in Salp chains II Physiologyrdquo Proceedings of theRoyal Society of London Series B Biological Sciences vol 210no 1181 pp 559ndash574 1980

[41] H T Ibrahim W J Mazher O N Ucan and O BayatldquoFeature selection using salp swarm algorithm for real bio-medical datasetsrdquo IJCSNS vol 17 no 12 pp 13ndash20 2017

12 Journal of Computer Networks and Communications

the number of variables inside the presented problemrepresented by n Accordingly a two-dimensional matrixdescribed as x will reservation the position of all Salps It issupposed too in search space there is ldquoFrdquo which is a foodsource as the target of the swarm +e following equation issuggested to upgrade the leader location [37]

x1j

Fj + c1 ubj minus lbj1113872 1113873c2 + lbj1113872 1113873 c3 ge 0

Fj minus c1 ubj minus lbj1113872 1113873c2 + lbj1113872 1113873 c3 lt 0

⎧⎪⎨

⎪⎩(10)

where x1j shows the position of the first Salp (leader) in

the jth dimension Fj is the position of the food source in thejth dimension ubj indicates the upper bound of jth di-mension lbj indicates the lower bound of jth dimension andc1 cthe2 and c3 are random numbers

Equation (10) shows that the leader only updates itsposition with respect to the food source +e coefficient c1 isthe most important parameter in SSA because it balancesexploration and exploitation defined as follows [37]

c1 2eminus(4lL)2

(11)

where l is the current iteration and L is the maximumnumber of iterations

+e parameters c2 and c3 are random numbers uni-formly generated in the interval of [0 1] In fact they dictateif the next position in jth dimension should be towardspositive infinity or negative infinity as well as the step size Toupdate the position of the followers the following equationsis utilized (Newtonrsquos law of motion) [37]

xij

12

at2

+ v0t (12)

where ige 2 xij shows the position of ith follower Salp in jth

dimension t is time v0 is the initial speed and a vfinalv0where v x minus x0t

Because the time in optimization is an iteration thediscrepancy between iterations is equal to 1 and consideringv0 0 this equation can be expressed as follows [37]

xij

12

xij + x

iminus1j1113872 1113873 (13)

where ige 2 and xij shows the position of ith follower Salp in

jth dimension Using equations (1) and (4) the Salp chainscan be simulated Figure 2 shows the pseudocode to im-plement the SSA algorithm

4 Formulation of WSN Localization Problem

WSN node localization problem is formulated using thesingle hop range-based distribution technique to estimatethe position of the unknown node coordinates (X Y) withthe aid of anchor nodes (position of known nodes) co-ordinates (x y) Anchor nodes are provided with a GPSdevice so it has the capability of automatically determiningits position Most of the nodes in the WSN are not equippedwith GPS due to high cost To measure the coordinates of Nunknown nodes the procedure followed is given below

Step 1 Randomly initialize the N unknown nodes and Manchor nodes within the communication range (R) Anchornodes measure their position and communicate their co-ordinates to their neighbors For all iterations the nodewhich settles at the end is termed as the reference node andthis node will act as anchor node

Step 2 +ree or more anchor nodes within the commu-nication range of a node are considered as a localized node

Step 3 Let (x y) be the coordinates of the target node to bedetermined and di be the distance between the target nodeand the ith anchor node

di

xminus xi( 11138572

+ yminusyi( 11138572

1113969

(14)

Step 4 +e optimization problem is formulated to minimizethe error of the localization problem +e objective functionfor the localization problem is formulated as

(a) (b)

Figure 1 (a) Individual Salp (b) Salps chain [37]

6 Journal of Computer Networks and Communications

f(x y) min sumM

i1

xminus xi( )2 + yminusyi( )2

radic( )

2 (15)

whereM is the anchor nodes within the transmission range(R) of the target node

Step 5 All target localized nodes (NL) are determined thewhole localization error is calculated as the mean of thesquare of distances of the estimated coordinates node (xi yi)and the actual node coordinates (Xi Yi) for i 1 2 NL

EL 1NL

sumL

i1

xi minusXi( )2 + yi minusYi( )2

radic( ) (16)

e performance of SSA algorithm evaluated using ELand the number of nonlocalized nodes NNL whereNNL [NminusNL]

Step 6 Repeat the steps 2ndash5 until all unknowntarget nodesget localized or no more nodes can be localized

5 Experimental Analysis

In this section the proposed WSN localization approach isevaluated under dierent scenarios and its performance iscompared to four other swarm-based algorithms (PSOBOA FA and GWO) in terms of localization accuracy andcomputing time e computations of the dierent algo-rithms are performed using MATLAB R2012b on a machineof Intel Core i7 CPU 4GB RAM and Windows7 operatingsystem e parametersrsquo values of the deployment area areshown in Table 1

For BOA the sensory modality c is set 001 whereasthe initial value of power exponent a is set to 01 [25] ForPSO initial values of ω 07 and c1 c2 1494 were rec-ommended for faster convergence after experimental tests[25] For FA the randomization parameter α is set to 025the absorption coecient c is set to 10 and the initial

attractiveness parameter β is set to 1 For GWO the pa-rameter a linearly decreases in the interval of [2 to 0] and theC parameter linearly increases from 0 to 2 [6] Finally forSSA c1 is calculated using equation (2) whereas c2 07 andc3 03 and c2 c3 are random numbers uniformly generatedfrom the interval [0 1]

51 Sensor Nodes Localization Using SSA In all conductedexperiments the coordinates of central nodes (N) anddestination nodes (M) are randomly congured during theconstruction of the deployment area e deployment areaincludes three types of nodes anchor nodes whose knownposition target nodes whose unknown position and lo-calized nodes whose positions are already estimated In thissection the performance of the SSA-based localization al-gorithm is evaluated under dierent scenarios using dif-ferent numbers of target nodes and dierent numbers ofanchors as shown in Figure 3

52 Comparison among Dierent Localization AlgorithmsIn this section SSA and the other swarm algorithms havebeen evaluated under dierent scenarios in terms of local-ization error computing time and number of localizednodes e obtained results of the dierent algorithms areshown in Table 2

Under the dierent scenarios (number of nodesnumberof anchors) it is noticed that for all the localization algo-rithms increasing the number of iterations increases both ofthe number of localized nodes and the computing time whilereduces the localization error is notice is rational because

Initialize the Salp population xi (i = 1 2 hellip n) considering ub and lbWhile (end condition is not satisfied)Calculate the fitness of each search agent (salp)F = the best search agentUpdate c1 by equation (2)

Update the position of the leading Salp by equation (1)Else

Update the position of the follower Salp by equation (4)End

End

EndReturn F

Amend the Salps based on the upper and lower bounds of variables

If (i == 1)For each Salp (xi)

Figure 2 Pseudocode of the SSA algorithm

Table 1 Parameters setting of simulation environment

Parameters ValuesSensor nodes Varies on sum6

i1ilowast 25Anchor nodes Varies on increment i i + 5Node transmission range (R) 30mDeployment area 100mtimes 100mMaximum number of iterations 100

Journal of Computer Networks and Communications 7

increasing the number of iterations means higher amountsof computations which requires longer computation timeOn the other side increasing the number of iterations meansthat the chance to nd a better solution get bigger hence thenumber of localized nodes get larger and the value of lo-calization error get smaller For better results analysis theexperimental results are summarized in Table 3

Based on Table 3 regarding the mean localization error(EL) there is no clear pattern that can be detected to rep-resent the relationship between this performance metric andthe number of target nodes and anchor nodes However it isnoticed that SSA has the best results regarding this per-formance metric compared to PSO BOA FA and GWOparticularly when the numbers of target nodes and anchor

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(a)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(b)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(c)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(d)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(e)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(f )

FIGURE 3 Node localization using dierent numbers of target nodes and anchor nodes (a) Target nodes 25 anchor nodes 10 (b) Targetnodes 50 anchor nodes 15 (c) Target nodes 75 anchor nodes 20 (d) Target nodes 100 anchor nodes 25 (e) Target nodes 125anchor nodes 30 (f ) Target nodes 150 anchor nodes 35

8 Journal of Computer Networks and Communications

nodes are increased Regarding the computing time it isnoticed that increasing the number of target nodes andanchor nodes increase the computing time for all local-ization algorithms However once again SSA has the bestcomputing time compared to PSO BOA FA and GWOFinally regarding the number of localized nodes (NL) it isnoticed that SSA has the best results compared to otherlocalization algorithms In addition it is noticed that SSAhas inferior EL when the percentage of (number of anchornodesnumber of target nodes) becomes larger such as(1025) in the first case in Table 3 +at is because thelocalization accuracy increases when the anchor densityincreases with respect to the number of the target nodes[25]

+e graphical representations of the experimentalresults for the different performance metrics are shown inFigures 4ndash6

6 Conclusion

Accurate node localization concerns many applications thatadopt WSNs In this paper a node localization algorithm hasbeen proposed based on a novel bioinspired algorithm calledSalp Swarm Algorithm (SSA) which handled the node lo-calization problem as an optimization problem+e proposedalgorithm has been implemented and validated in differentWSN deployments using different numbers of target nodesand anchor nodes Moreover the proposed algorithm hasbeen evaluated and compared to four well-known optimi-zation algorithms namely PSO BOA FA and GWO interms of localization accuracy computing time and severallocalized nodes +e obtained simulation results have provedthe superiority of the proposed algorithm compared to theother localization algorithms regarding the different perfor-mance metrics In the future work the proposed approach

Table 2 Performance metrics of different localization algorithms

Targetnodes

Anchornodes

No ofiterations

PSO BOA FA GWO SSAEL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10

25 0818 040 16 0232 037 22 0265 06 19 0744 022 20 0465 035 2250 0812 040 15 0225 038 23 0262 12 20 0741 041 21 0462 035 2375 0803 041 18 0223 039 25 0258 15 21 0741 054 21 0458 036 23100 0792 041 18 0221 040 25 0251 18 19 0740 079 23 0451 037 24

50 15

25 0419 071 41 0338 084 47 0477 16 46 0690 042 44 0477 067 4350 0426 073 47 0332 085 48 0473 19 49 0688 063 45 0472 069 4775 0429 076 46 0326 086 49 0465 25 49 0686 081 46 0468 069 48100 0434 076 48 0323 086 49 0465 35 48 0682 098 48 0464 070 50

75 20

25 0735 131 73 0257 149 66 0519 29 71 0641 072 72 0519 090 6950 0728 132 74 0257 149 66 0513 38 72 0641 095 72 0513 092 7275 0728 133 74 0255 150 70 0504 47 73 0638 13 73 0504 095 73100 0724 135 75 0253 152 72 0503 52 73 0635 14 74 0503 096 75

100 25

25 0661 210 97 0355 240 97 0711 38 98 0611 11 95 0511 131 9850 0658 216 97 0355 244 99 0709 42 98 0606 15 97 0509 133 9875 0642 217 99 0333 247 100 0702 56 99 0602 18 98 0502 136 99100 0641 220 100 0331 250 100 0704 63 98 0602 21 98 0504 137 100

125 30

25 0754 487 120 0549 384 122 0829 27 122 0589 15 122 0529 167 12350 0748 486 121 0548 385 123 0824 45 123 0580 22 123 0524 168 12475 0750 489 122 0534 386 124 0822 59 125 0580 28 123 0522 170 125100 0752 495 125 0534 389 124 0822 65 124 0572 33 125 0522 172 125

150 35

25 0625 541 145 0766 588 147 0911 25 149 0559 28 148 0511 212 14950 0622 542 146 0765 561 148 0909 42 150 0547 36 149 0509 214 14975 0619 544 148 0763 564 149 0904 64 150 0523 43 150 0504 216 150100 0616 545 150 0763 569 149 0904 72 149 0523 48 150 0504 218 150

Table 3 Summary of experimental results of the different localization algorithms

Target nodes Anchor nodesPSO BOA FA GWO SSA

EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10 079 040 18 022 038 25 025 18 19 074 054 21 045 035 2450 15 043 076 46 032 086 49 042 25 49 069 081 46 046 069 4875 20 072 135 75 025 152 68 051 47 73 064 095 72 050 096 73100 25 065 216 100 035 245 100 070 53 98 060 21 98 051 135 100125 30 074 490 123 054 387 124 082 65 124 058 28 123 052 170 125150 35 062 543 149 076 565 149 090 72 149 052 43 150 050 215 150

Journal of Computer Networks and Communications 9

0

01

02

03

04M

ean

loca

lizat

ion

erro

r (m

)05

06

07

08

09

1

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

Figure 4 e localization error of the dierent localization algorithms in dierent WSN deployments

0

1

2

3

4

5

6

7

8

9

10

Com

putin

g tim

e (S)

(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35

Figure 5 e computing time of the dierent localization algorithms in dierent WSN deployments

0 20 40 60 80 100 120 140Number of localized nodes

(Num

ber o

f tar

get n

odes

ndashnu

mbe

r of

anch

ors)

SSAGWOFA

BOAPSO

25 ndash 10

50 ndash 15

75 ndash 20

100 ndash 25

125 ndash 30

150 ndash 35

Figure 6 e number of localized nodes of the dierent localization algorithms in dierent WSN deployments

10 Journal of Computer Networks and Communications

can be hybridized with other algorithm to reduce the local-ization error

Data Availability

No data were used to support this study Because our articlediscusses the problem of localization in WSNs our exper-imental results have been applied based on several networksrsquosize

Conflicts of Interest

+e authors declare that they have no conflicts of interest

References

[1] R V Kulkarni G K Venayagamoorthy and M X ChengldquoBio-inspired node localization in wireless sensor networksrdquoin Proceedings of IEEE International Conference in SystemsMan and Cybernetics (SMC) pp 205ndash210 San Antonio TXUSA October 2009

[2] D Lavanya and S K Udgata ldquoSwarm intelligence-basedlocalization in wireless sensor networksrdquo in InternationalWorkshop on Multi-Disciplinary Trends in Artificial In-telligence pp 317ndash328 Springer Berlin Germany 2011

[3] J Wang R K Ghosh and S K Das ldquoA survey on sensorlocalizationrdquo Journal of Control Aeory and Applicationsvol 8 no 1 pp 2ndash11 2010

[4] J Aspnes T Eren D K Goldenberg et al ldquoA theory ofnetwork localizationrdquo IEEE Transactions on Mobile Com-puting vol 5 no 12 pp 1663ndash1678 2006

[5] S Goyal and M S Patterh ldquoModified bat algorithm for lo-calization of wireless sensor networkrdquo Wireless PersonalCommunications vol 86 no 2 pp 657ndash670 2015

[6] M M Ahmed E H Houssein A E Hassanien A Taha andE Hassanien ldquoMaximizing lifetime of wireless sensor net-works based on whale optimization algorithmrdquo in In-ternational Conference on Advanced Intelligent Systems andInformatics pp 724ndash733 Springer Berlin Germany 2017

[7] K S Low H A Nguyen and H Guo ldquoA particle swarmoptimization approach for the localization of a wireless sensornetworkrdquo in Proceedings of IEEE International Symposium inIndustrial Electronics (ISIE) pp 1820ndash1825 VancouverCanada June 2008

[8] D Manjarres J Del Ser S Gil-Lopez M Vecchio I Landa-Torres and R Lopez-Valcarce ldquoOn the application of a hybridharmony search algorithm to node localization in anchor-based wireless sensor networksrdquo in Proceedings of IEEE In-ternational Conference in Intelligent Systems Design andApplications (ISDA) pp 1014ndash1019 Cordoba Spain No-vember 2011

[9] M Li W Xiong and Q Liang ldquoAn improved abc-based nodelocalization algorithm for wireless sensor networkrdquo in Pro-ceedings of IEEE International Conference in Wireless Com-munications Networking and Mobile Computing (WiCOM)pp 1ndash4 Barcelona Spain September 2012

[10] A Tamizharasi R Arthi and K Murugan ldquoBio-inspiredalgorithm for optimizing the localization of wireless sensornetworksrdquo in Proceedings of IEEE International Conference inComputing Communications and Networking Technologies(ICCCNT) pp 1ndash5 Tiruchengode India June 2013

[11] C Tang R Liu and J Ni ldquoA novel wireless sensor networklocalization approach localization based on plant growth

simulation algorithmrdquo Elektronika ir Elektrotechnika vol 19no 8 pp 97ndash100 2013

[12] O D Jegede and K Ferens ldquoA genetic algorithm for nodelocalization in wireless sensor networksrdquo in Proceedings ofInternational Conference on Genetic and EvolutionaryMethods (GEM) Ae Steering Committee of the World Con-gress in Computer Science Computer Engineering and AppliedComputing (WorldComp) pp 1ndash7 Las Vegas NV USA July2013

[13] S Goyal and M S Patterh ldquoWireless sensor network local-ization based on cuckoo search algorithmrdquo Wireless personalcommunications vol 79 no 1 pp 223ndash234 2014

[14] L Dan and W Xian-bin ldquoAn improved PSO algorithm fordistributed localization in wireless sensor networksrdquo In-ternational Journal of Distributed Sensor Networks vol 11no 7 Article ID 970272 2015

[15] R Krishnaprabha and A Gopakumar ldquoPerformance ofgravitational search algorithm in wireless sensor networklocalizationrdquo in Proceedings of IEEE National Conference inCommunication Signal Processing and Networking (NCCSN)pp 1ndash6 Palakkad India October 2014

[16] B Peng and L Li ldquoAn improved localization algorithm basedon genetic algorithm in wireless sensor networksrdquo CognitiveNeurodynamics vol 9 no 2 pp 249ndash256 2015

[17] V O Sai C S Shieh T T Nguyen Y C Lin M F Horngand Q D Le ldquoParallel firefly algorithm for localization al-gorithm in wireless sensor networkrdquo in Proceedings of IEEEInternational Conference in Robot Vision and Signal Pro-cessing (RVSP) pp 300ndash305 Kaohsiung China November2015

[18] S Sivakumar and R Venkatesan ldquoMeta-heuristic approachesfor minimizing error in localization of wireless sensor net-worksrdquo Applied Soft Computing vol 36 pp 506ndash518 2015

[19] Z Sun L Tao X Wang and Z Zhou ldquoLocalization algorithmin wireless sensor networks based on multiobjective particleswarm optimizationrdquo International Journal of DistributedSensor Networks vol 11 no 8 Article ID 716291 2015

[20] A Arsic M Tuba and M Jordanski ldquoFireworks algorithmapplied to wireless sensor networks localization problemrdquo inProceedings of IEEE Congress on Evolutionary Computation(CEC) pp 4038ndash4044 Vancouver Canada July 2016

[21] C S Shieh V O Sai Y C Lin T F Lee T T Nguyen andQ D Le ldquoImproved node localization for WSN using heu-ristic optimization approachesrdquo in Proceedings of IEEE In-ternational Conference in Networking and NetworkApplications (NaNA) pp 95ndash98 Hokkaido Japan July 2016

[22] J Cheng and L Xia ldquoAn effective Cuckoo search algorithm fornode localization in wireless sensor networkrdquo Sensors vol 16no 9 p 1390 2016

[23] P T Daely and S Y Shin ldquoRange based wireless node lo-calization using Dragonfly Algorithmrdquo in Proceedings of IEEEInternational Conference in Ubiquitous and Future Networks(ICUFN) pp 1012ndash1015 Vienna Austria July 2016

[24] T T Nguyen J S Pan S C Chu J F Roddick and T K DaoldquoOptimization localization in wireless sensor network basedon multi-objective firefly algorithmrdquo Journal of NetworkIntelligence vol 1 no 4 pp 130ndash138 2016

[25] S Arora and S Singh ldquoNode localization in wireless sensornetworks using butterfly optimization algorithmrdquo ArabianJournal for Science and Engineering vol 42 no 8pp 3325ndash3335 2017

[26] S R Sujatha and M Siddappa ldquoNode localization method forwireless sensor networks based on hybrid optimization ofparticle swarm optimization and differential evolutionrdquo IOSR

Journal of Computer Networks and Communications 11

Journal of Computer Engineering vol 19 no 2 pp 07ndash122017

[27] R Kaur and S Arora ldquonature inspired range based wirelesssensor node localization algorithmsrdquo International Journal ofInteractive Multimedia and Artificial Intelligence vol 4pp 7ndash17 2017

[28] A +arwat E H Houssein M M Ahmed A E Hassanienand T Gabel ldquoMOGOA algorithm for constrained and un-constrained multi-objective optimization problemsrdquo AppliedIntelligence vol 48 no 8 pp 2268ndash2283 2017

[29] A A Ewees M Abd Elaziz and E H Houssein ldquoImprovedgrasshopper optimization algorithm using opposition-basedlearningrdquo Expert Systems with Applications vol 112pp 156ndash172 2018

[30] A G Hussien E H Houssein and A E Hassanien ldquoA binarywhale optimization algorithm with hyperbolic tangent fitnessfunction for feature selectionrdquo in Proceedings of Eighth In-ternational Conference on Intelligent Computing and In-formation Systems (ICICIS) pp 166ndash172 Cairo EgyptDecember 2017

[31] B Xue M Zhang and W N Browne ldquoParticle swarm op-timization for feature selection in classification a multi-objective approachrdquo IEEE Transactions on Cyberneticsvol 43 no 6 pp 1656ndash1671 2013

[32] S Arora and S Singh ldquoButterfly algorithmwith levy flights forglobal optimizationrdquo in Proceedings of IEEE InternationalConference in Signal Processing Computing and Control(ISPCC) pp 220ndash224 Waknaghat India September 2015

[33] S Arora and S Singh ldquoAn improved butterfly optimizationalgorithm with chaosrdquo Journal of Intelligent and Fuzzy Sys-tems vol 32 no 1 pp 1079ndash1088 2017

[34] R B Blair and A E Launer ldquoButterfly diversity and humanland use species assemblages along an urban grandientrdquoBiological Conservation vol 80 no 1 pp 113ndash125 1997

[35] X S Yang and X He ldquoFirefly algorithm recent advances andapplicationsrdquo 2013 httpsarxivorgabs13083898

[36] X S Yang ldquoFirefly algorithms for multimodal optimizationrdquoin Proceedings of International Symposium on Stochastic Al-gorithms pp 169ndash178 Sapporo Japan October 2009

[37] S Mirjalili S M Mirjalili and A Lewis ldquoGrey wolf opti-mizerrdquo Advances in Engineering Software vol 69 pp 46ndash612014

[38] L P Madin ldquoAspects of jet propulsion in Salpsrdquo CanadianJournal of Zoology vol 68 no 4 pp 765ndash777 1990

[39] S Mirjalili A H Gandomi S Z Mirjalili S Saremi H Farisand S M Mirjalili ldquoSalp swarm algorithm a bio-inspiredoptimizer for engineering design problemsrdquo Advances inEngineering Software vol 114 pp 163ndash191 2017

[40] P A Anderson and Q Bone ldquoCommunication betweenindividuals in Salp chains II Physiologyrdquo Proceedings of theRoyal Society of London Series B Biological Sciences vol 210no 1181 pp 559ndash574 1980

[41] H T Ibrahim W J Mazher O N Ucan and O BayatldquoFeature selection using salp swarm algorithm for real bio-medical datasetsrdquo IJCSNS vol 17 no 12 pp 13ndash20 2017

12 Journal of Computer Networks and Communications

f(x y) min sumM

i1

xminus xi( )2 + yminusyi( )2

radic( )

2 (15)

whereM is the anchor nodes within the transmission range(R) of the target node

Step 5 All target localized nodes (NL) are determined thewhole localization error is calculated as the mean of thesquare of distances of the estimated coordinates node (xi yi)and the actual node coordinates (Xi Yi) for i 1 2 NL

EL 1NL

sumL

i1

xi minusXi( )2 + yi minusYi( )2

radic( ) (16)

e performance of SSA algorithm evaluated using ELand the number of nonlocalized nodes NNL whereNNL [NminusNL]

Step 6 Repeat the steps 2ndash5 until all unknowntarget nodesget localized or no more nodes can be localized

5 Experimental Analysis

In this section the proposed WSN localization approach isevaluated under dierent scenarios and its performance iscompared to four other swarm-based algorithms (PSOBOA FA and GWO) in terms of localization accuracy andcomputing time e computations of the dierent algo-rithms are performed using MATLAB R2012b on a machineof Intel Core i7 CPU 4GB RAM and Windows7 operatingsystem e parametersrsquo values of the deployment area areshown in Table 1

For BOA the sensory modality c is set 001 whereasthe initial value of power exponent a is set to 01 [25] ForPSO initial values of ω 07 and c1 c2 1494 were rec-ommended for faster convergence after experimental tests[25] For FA the randomization parameter α is set to 025the absorption coecient c is set to 10 and the initial

attractiveness parameter β is set to 1 For GWO the pa-rameter a linearly decreases in the interval of [2 to 0] and theC parameter linearly increases from 0 to 2 [6] Finally forSSA c1 is calculated using equation (2) whereas c2 07 andc3 03 and c2 c3 are random numbers uniformly generatedfrom the interval [0 1]

51 Sensor Nodes Localization Using SSA In all conductedexperiments the coordinates of central nodes (N) anddestination nodes (M) are randomly congured during theconstruction of the deployment area e deployment areaincludes three types of nodes anchor nodes whose knownposition target nodes whose unknown position and lo-calized nodes whose positions are already estimated In thissection the performance of the SSA-based localization al-gorithm is evaluated under dierent scenarios using dif-ferent numbers of target nodes and dierent numbers ofanchors as shown in Figure 3

52 Comparison among Dierent Localization AlgorithmsIn this section SSA and the other swarm algorithms havebeen evaluated under dierent scenarios in terms of local-ization error computing time and number of localizednodes e obtained results of the dierent algorithms areshown in Table 2

Under the dierent scenarios (number of nodesnumberof anchors) it is noticed that for all the localization algo-rithms increasing the number of iterations increases both ofthe number of localized nodes and the computing time whilereduces the localization error is notice is rational because

Initialize the Salp population xi (i = 1 2 hellip n) considering ub and lbWhile (end condition is not satisfied)Calculate the fitness of each search agent (salp)F = the best search agentUpdate c1 by equation (2)

Update the position of the leading Salp by equation (1)Else

Update the position of the follower Salp by equation (4)End

End

EndReturn F

Amend the Salps based on the upper and lower bounds of variables

If (i == 1)For each Salp (xi)

Figure 2 Pseudocode of the SSA algorithm

Table 1 Parameters setting of simulation environment

Parameters ValuesSensor nodes Varies on sum6

i1ilowast 25Anchor nodes Varies on increment i i + 5Node transmission range (R) 30mDeployment area 100mtimes 100mMaximum number of iterations 100

Journal of Computer Networks and Communications 7

increasing the number of iterations means higher amountsof computations which requires longer computation timeOn the other side increasing the number of iterations meansthat the chance to nd a better solution get bigger hence thenumber of localized nodes get larger and the value of lo-calization error get smaller For better results analysis theexperimental results are summarized in Table 3

Based on Table 3 regarding the mean localization error(EL) there is no clear pattern that can be detected to rep-resent the relationship between this performance metric andthe number of target nodes and anchor nodes However it isnoticed that SSA has the best results regarding this per-formance metric compared to PSO BOA FA and GWOparticularly when the numbers of target nodes and anchor

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(a)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(b)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(c)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(d)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(e)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(f )

FIGURE 3 Node localization using dierent numbers of target nodes and anchor nodes (a) Target nodes 25 anchor nodes 10 (b) Targetnodes 50 anchor nodes 15 (c) Target nodes 75 anchor nodes 20 (d) Target nodes 100 anchor nodes 25 (e) Target nodes 125anchor nodes 30 (f ) Target nodes 150 anchor nodes 35

8 Journal of Computer Networks and Communications

nodes are increased Regarding the computing time it isnoticed that increasing the number of target nodes andanchor nodes increase the computing time for all local-ization algorithms However once again SSA has the bestcomputing time compared to PSO BOA FA and GWOFinally regarding the number of localized nodes (NL) it isnoticed that SSA has the best results compared to otherlocalization algorithms In addition it is noticed that SSAhas inferior EL when the percentage of (number of anchornodesnumber of target nodes) becomes larger such as(1025) in the first case in Table 3 +at is because thelocalization accuracy increases when the anchor densityincreases with respect to the number of the target nodes[25]

+e graphical representations of the experimentalresults for the different performance metrics are shown inFigures 4ndash6

6 Conclusion

Accurate node localization concerns many applications thatadopt WSNs In this paper a node localization algorithm hasbeen proposed based on a novel bioinspired algorithm calledSalp Swarm Algorithm (SSA) which handled the node lo-calization problem as an optimization problem+e proposedalgorithm has been implemented and validated in differentWSN deployments using different numbers of target nodesand anchor nodes Moreover the proposed algorithm hasbeen evaluated and compared to four well-known optimi-zation algorithms namely PSO BOA FA and GWO interms of localization accuracy computing time and severallocalized nodes +e obtained simulation results have provedthe superiority of the proposed algorithm compared to theother localization algorithms regarding the different perfor-mance metrics In the future work the proposed approach

Table 2 Performance metrics of different localization algorithms

Targetnodes

Anchornodes

No ofiterations

PSO BOA FA GWO SSAEL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10

25 0818 040 16 0232 037 22 0265 06 19 0744 022 20 0465 035 2250 0812 040 15 0225 038 23 0262 12 20 0741 041 21 0462 035 2375 0803 041 18 0223 039 25 0258 15 21 0741 054 21 0458 036 23100 0792 041 18 0221 040 25 0251 18 19 0740 079 23 0451 037 24

50 15

25 0419 071 41 0338 084 47 0477 16 46 0690 042 44 0477 067 4350 0426 073 47 0332 085 48 0473 19 49 0688 063 45 0472 069 4775 0429 076 46 0326 086 49 0465 25 49 0686 081 46 0468 069 48100 0434 076 48 0323 086 49 0465 35 48 0682 098 48 0464 070 50

75 20

25 0735 131 73 0257 149 66 0519 29 71 0641 072 72 0519 090 6950 0728 132 74 0257 149 66 0513 38 72 0641 095 72 0513 092 7275 0728 133 74 0255 150 70 0504 47 73 0638 13 73 0504 095 73100 0724 135 75 0253 152 72 0503 52 73 0635 14 74 0503 096 75

100 25

25 0661 210 97 0355 240 97 0711 38 98 0611 11 95 0511 131 9850 0658 216 97 0355 244 99 0709 42 98 0606 15 97 0509 133 9875 0642 217 99 0333 247 100 0702 56 99 0602 18 98 0502 136 99100 0641 220 100 0331 250 100 0704 63 98 0602 21 98 0504 137 100

125 30

25 0754 487 120 0549 384 122 0829 27 122 0589 15 122 0529 167 12350 0748 486 121 0548 385 123 0824 45 123 0580 22 123 0524 168 12475 0750 489 122 0534 386 124 0822 59 125 0580 28 123 0522 170 125100 0752 495 125 0534 389 124 0822 65 124 0572 33 125 0522 172 125

150 35

25 0625 541 145 0766 588 147 0911 25 149 0559 28 148 0511 212 14950 0622 542 146 0765 561 148 0909 42 150 0547 36 149 0509 214 14975 0619 544 148 0763 564 149 0904 64 150 0523 43 150 0504 216 150100 0616 545 150 0763 569 149 0904 72 149 0523 48 150 0504 218 150

Table 3 Summary of experimental results of the different localization algorithms

Target nodes Anchor nodesPSO BOA FA GWO SSA

EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10 079 040 18 022 038 25 025 18 19 074 054 21 045 035 2450 15 043 076 46 032 086 49 042 25 49 069 081 46 046 069 4875 20 072 135 75 025 152 68 051 47 73 064 095 72 050 096 73100 25 065 216 100 035 245 100 070 53 98 060 21 98 051 135 100125 30 074 490 123 054 387 124 082 65 124 058 28 123 052 170 125150 35 062 543 149 076 565 149 090 72 149 052 43 150 050 215 150

Journal of Computer Networks and Communications 9

0

01

02

03

04M

ean

loca

lizat

ion

erro

r (m

)05

06

07

08

09

1

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

Figure 4 e localization error of the dierent localization algorithms in dierent WSN deployments

0

1

2

3

4

5

6

7

8

9

10

Com

putin

g tim

e (S)

(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35

Figure 5 e computing time of the dierent localization algorithms in dierent WSN deployments

0 20 40 60 80 100 120 140Number of localized nodes

(Num

ber o

f tar

get n

odes

ndashnu

mbe

r of

anch

ors)

SSAGWOFA

BOAPSO

25 ndash 10

50 ndash 15

75 ndash 20

100 ndash 25

125 ndash 30

150 ndash 35

Figure 6 e number of localized nodes of the dierent localization algorithms in dierent WSN deployments

10 Journal of Computer Networks and Communications

can be hybridized with other algorithm to reduce the local-ization error

Data Availability

No data were used to support this study Because our articlediscusses the problem of localization in WSNs our exper-imental results have been applied based on several networksrsquosize

Conflicts of Interest

+e authors declare that they have no conflicts of interest

References

[1] R V Kulkarni G K Venayagamoorthy and M X ChengldquoBio-inspired node localization in wireless sensor networksrdquoin Proceedings of IEEE International Conference in SystemsMan and Cybernetics (SMC) pp 205ndash210 San Antonio TXUSA October 2009

[2] D Lavanya and S K Udgata ldquoSwarm intelligence-basedlocalization in wireless sensor networksrdquo in InternationalWorkshop on Multi-Disciplinary Trends in Artificial In-telligence pp 317ndash328 Springer Berlin Germany 2011

[3] J Wang R K Ghosh and S K Das ldquoA survey on sensorlocalizationrdquo Journal of Control Aeory and Applicationsvol 8 no 1 pp 2ndash11 2010

[4] J Aspnes T Eren D K Goldenberg et al ldquoA theory ofnetwork localizationrdquo IEEE Transactions on Mobile Com-puting vol 5 no 12 pp 1663ndash1678 2006

[5] S Goyal and M S Patterh ldquoModified bat algorithm for lo-calization of wireless sensor networkrdquo Wireless PersonalCommunications vol 86 no 2 pp 657ndash670 2015

[6] M M Ahmed E H Houssein A E Hassanien A Taha andE Hassanien ldquoMaximizing lifetime of wireless sensor net-works based on whale optimization algorithmrdquo in In-ternational Conference on Advanced Intelligent Systems andInformatics pp 724ndash733 Springer Berlin Germany 2017

[7] K S Low H A Nguyen and H Guo ldquoA particle swarmoptimization approach for the localization of a wireless sensornetworkrdquo in Proceedings of IEEE International Symposium inIndustrial Electronics (ISIE) pp 1820ndash1825 VancouverCanada June 2008

[8] D Manjarres J Del Ser S Gil-Lopez M Vecchio I Landa-Torres and R Lopez-Valcarce ldquoOn the application of a hybridharmony search algorithm to node localization in anchor-based wireless sensor networksrdquo in Proceedings of IEEE In-ternational Conference in Intelligent Systems Design andApplications (ISDA) pp 1014ndash1019 Cordoba Spain No-vember 2011

[9] M Li W Xiong and Q Liang ldquoAn improved abc-based nodelocalization algorithm for wireless sensor networkrdquo in Pro-ceedings of IEEE International Conference in Wireless Com-munications Networking and Mobile Computing (WiCOM)pp 1ndash4 Barcelona Spain September 2012

[10] A Tamizharasi R Arthi and K Murugan ldquoBio-inspiredalgorithm for optimizing the localization of wireless sensornetworksrdquo in Proceedings of IEEE International Conference inComputing Communications and Networking Technologies(ICCCNT) pp 1ndash5 Tiruchengode India June 2013

[11] C Tang R Liu and J Ni ldquoA novel wireless sensor networklocalization approach localization based on plant growth

simulation algorithmrdquo Elektronika ir Elektrotechnika vol 19no 8 pp 97ndash100 2013

[12] O D Jegede and K Ferens ldquoA genetic algorithm for nodelocalization in wireless sensor networksrdquo in Proceedings ofInternational Conference on Genetic and EvolutionaryMethods (GEM) Ae Steering Committee of the World Con-gress in Computer Science Computer Engineering and AppliedComputing (WorldComp) pp 1ndash7 Las Vegas NV USA July2013

[13] S Goyal and M S Patterh ldquoWireless sensor network local-ization based on cuckoo search algorithmrdquo Wireless personalcommunications vol 79 no 1 pp 223ndash234 2014

[14] L Dan and W Xian-bin ldquoAn improved PSO algorithm fordistributed localization in wireless sensor networksrdquo In-ternational Journal of Distributed Sensor Networks vol 11no 7 Article ID 970272 2015

[15] R Krishnaprabha and A Gopakumar ldquoPerformance ofgravitational search algorithm in wireless sensor networklocalizationrdquo in Proceedings of IEEE National Conference inCommunication Signal Processing and Networking (NCCSN)pp 1ndash6 Palakkad India October 2014

[16] B Peng and L Li ldquoAn improved localization algorithm basedon genetic algorithm in wireless sensor networksrdquo CognitiveNeurodynamics vol 9 no 2 pp 249ndash256 2015

[17] V O Sai C S Shieh T T Nguyen Y C Lin M F Horngand Q D Le ldquoParallel firefly algorithm for localization al-gorithm in wireless sensor networkrdquo in Proceedings of IEEEInternational Conference in Robot Vision and Signal Pro-cessing (RVSP) pp 300ndash305 Kaohsiung China November2015

[18] S Sivakumar and R Venkatesan ldquoMeta-heuristic approachesfor minimizing error in localization of wireless sensor net-worksrdquo Applied Soft Computing vol 36 pp 506ndash518 2015

[19] Z Sun L Tao X Wang and Z Zhou ldquoLocalization algorithmin wireless sensor networks based on multiobjective particleswarm optimizationrdquo International Journal of DistributedSensor Networks vol 11 no 8 Article ID 716291 2015

[20] A Arsic M Tuba and M Jordanski ldquoFireworks algorithmapplied to wireless sensor networks localization problemrdquo inProceedings of IEEE Congress on Evolutionary Computation(CEC) pp 4038ndash4044 Vancouver Canada July 2016

[21] C S Shieh V O Sai Y C Lin T F Lee T T Nguyen andQ D Le ldquoImproved node localization for WSN using heu-ristic optimization approachesrdquo in Proceedings of IEEE In-ternational Conference in Networking and NetworkApplications (NaNA) pp 95ndash98 Hokkaido Japan July 2016

[22] J Cheng and L Xia ldquoAn effective Cuckoo search algorithm fornode localization in wireless sensor networkrdquo Sensors vol 16no 9 p 1390 2016

[23] P T Daely and S Y Shin ldquoRange based wireless node lo-calization using Dragonfly Algorithmrdquo in Proceedings of IEEEInternational Conference in Ubiquitous and Future Networks(ICUFN) pp 1012ndash1015 Vienna Austria July 2016

[24] T T Nguyen J S Pan S C Chu J F Roddick and T K DaoldquoOptimization localization in wireless sensor network basedon multi-objective firefly algorithmrdquo Journal of NetworkIntelligence vol 1 no 4 pp 130ndash138 2016

[25] S Arora and S Singh ldquoNode localization in wireless sensornetworks using butterfly optimization algorithmrdquo ArabianJournal for Science and Engineering vol 42 no 8pp 3325ndash3335 2017

[26] S R Sujatha and M Siddappa ldquoNode localization method forwireless sensor networks based on hybrid optimization ofparticle swarm optimization and differential evolutionrdquo IOSR

Journal of Computer Networks and Communications 11

Journal of Computer Engineering vol 19 no 2 pp 07ndash122017

[27] R Kaur and S Arora ldquonature inspired range based wirelesssensor node localization algorithmsrdquo International Journal ofInteractive Multimedia and Artificial Intelligence vol 4pp 7ndash17 2017

[28] A +arwat E H Houssein M M Ahmed A E Hassanienand T Gabel ldquoMOGOA algorithm for constrained and un-constrained multi-objective optimization problemsrdquo AppliedIntelligence vol 48 no 8 pp 2268ndash2283 2017

[29] A A Ewees M Abd Elaziz and E H Houssein ldquoImprovedgrasshopper optimization algorithm using opposition-basedlearningrdquo Expert Systems with Applications vol 112pp 156ndash172 2018

[30] A G Hussien E H Houssein and A E Hassanien ldquoA binarywhale optimization algorithm with hyperbolic tangent fitnessfunction for feature selectionrdquo in Proceedings of Eighth In-ternational Conference on Intelligent Computing and In-formation Systems (ICICIS) pp 166ndash172 Cairo EgyptDecember 2017

[31] B Xue M Zhang and W N Browne ldquoParticle swarm op-timization for feature selection in classification a multi-objective approachrdquo IEEE Transactions on Cyberneticsvol 43 no 6 pp 1656ndash1671 2013

[32] S Arora and S Singh ldquoButterfly algorithmwith levy flights forglobal optimizationrdquo in Proceedings of IEEE InternationalConference in Signal Processing Computing and Control(ISPCC) pp 220ndash224 Waknaghat India September 2015

[33] S Arora and S Singh ldquoAn improved butterfly optimizationalgorithm with chaosrdquo Journal of Intelligent and Fuzzy Sys-tems vol 32 no 1 pp 1079ndash1088 2017

[34] R B Blair and A E Launer ldquoButterfly diversity and humanland use species assemblages along an urban grandientrdquoBiological Conservation vol 80 no 1 pp 113ndash125 1997

[35] X S Yang and X He ldquoFirefly algorithm recent advances andapplicationsrdquo 2013 httpsarxivorgabs13083898

[36] X S Yang ldquoFirefly algorithms for multimodal optimizationrdquoin Proceedings of International Symposium on Stochastic Al-gorithms pp 169ndash178 Sapporo Japan October 2009

[37] S Mirjalili S M Mirjalili and A Lewis ldquoGrey wolf opti-mizerrdquo Advances in Engineering Software vol 69 pp 46ndash612014

[38] L P Madin ldquoAspects of jet propulsion in Salpsrdquo CanadianJournal of Zoology vol 68 no 4 pp 765ndash777 1990

[39] S Mirjalili A H Gandomi S Z Mirjalili S Saremi H Farisand S M Mirjalili ldquoSalp swarm algorithm a bio-inspiredoptimizer for engineering design problemsrdquo Advances inEngineering Software vol 114 pp 163ndash191 2017

[40] P A Anderson and Q Bone ldquoCommunication betweenindividuals in Salp chains II Physiologyrdquo Proceedings of theRoyal Society of London Series B Biological Sciences vol 210no 1181 pp 559ndash574 1980

[41] H T Ibrahim W J Mazher O N Ucan and O BayatldquoFeature selection using salp swarm algorithm for real bio-medical datasetsrdquo IJCSNS vol 17 no 12 pp 13ndash20 2017

12 Journal of Computer Networks and Communications

increasing the number of iterations means higher amountsof computations which requires longer computation timeOn the other side increasing the number of iterations meansthat the chance to nd a better solution get bigger hence thenumber of localized nodes get larger and the value of lo-calization error get smaller For better results analysis theexperimental results are summarized in Table 3

Based on Table 3 regarding the mean localization error(EL) there is no clear pattern that can be detected to rep-resent the relationship between this performance metric andthe number of target nodes and anchor nodes However it isnoticed that SSA has the best results regarding this per-formance metric compared to PSO BOA FA and GWOparticularly when the numbers of target nodes and anchor

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(a)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(b)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(c)

20 40 60 80 1000

0

20

40

60

80

100 Anchor locationTarget locationEstimated node location

(d)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(e)

20 40 60 80 1000

0

20

40

60

80

100

Anchor locationTarget locationEstimated node location

(f )

FIGURE 3 Node localization using dierent numbers of target nodes and anchor nodes (a) Target nodes 25 anchor nodes 10 (b) Targetnodes 50 anchor nodes 15 (c) Target nodes 75 anchor nodes 20 (d) Target nodes 100 anchor nodes 25 (e) Target nodes 125anchor nodes 30 (f ) Target nodes 150 anchor nodes 35

8 Journal of Computer Networks and Communications

nodes are increased Regarding the computing time it isnoticed that increasing the number of target nodes andanchor nodes increase the computing time for all local-ization algorithms However once again SSA has the bestcomputing time compared to PSO BOA FA and GWOFinally regarding the number of localized nodes (NL) it isnoticed that SSA has the best results compared to otherlocalization algorithms In addition it is noticed that SSAhas inferior EL when the percentage of (number of anchornodesnumber of target nodes) becomes larger such as(1025) in the first case in Table 3 +at is because thelocalization accuracy increases when the anchor densityincreases with respect to the number of the target nodes[25]

+e graphical representations of the experimentalresults for the different performance metrics are shown inFigures 4ndash6

6 Conclusion

Accurate node localization concerns many applications thatadopt WSNs In this paper a node localization algorithm hasbeen proposed based on a novel bioinspired algorithm calledSalp Swarm Algorithm (SSA) which handled the node lo-calization problem as an optimization problem+e proposedalgorithm has been implemented and validated in differentWSN deployments using different numbers of target nodesand anchor nodes Moreover the proposed algorithm hasbeen evaluated and compared to four well-known optimi-zation algorithms namely PSO BOA FA and GWO interms of localization accuracy computing time and severallocalized nodes +e obtained simulation results have provedthe superiority of the proposed algorithm compared to theother localization algorithms regarding the different perfor-mance metrics In the future work the proposed approach

Table 2 Performance metrics of different localization algorithms

Targetnodes

Anchornodes

No ofiterations

PSO BOA FA GWO SSAEL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10

25 0818 040 16 0232 037 22 0265 06 19 0744 022 20 0465 035 2250 0812 040 15 0225 038 23 0262 12 20 0741 041 21 0462 035 2375 0803 041 18 0223 039 25 0258 15 21 0741 054 21 0458 036 23100 0792 041 18 0221 040 25 0251 18 19 0740 079 23 0451 037 24

50 15

25 0419 071 41 0338 084 47 0477 16 46 0690 042 44 0477 067 4350 0426 073 47 0332 085 48 0473 19 49 0688 063 45 0472 069 4775 0429 076 46 0326 086 49 0465 25 49 0686 081 46 0468 069 48100 0434 076 48 0323 086 49 0465 35 48 0682 098 48 0464 070 50

75 20

25 0735 131 73 0257 149 66 0519 29 71 0641 072 72 0519 090 6950 0728 132 74 0257 149 66 0513 38 72 0641 095 72 0513 092 7275 0728 133 74 0255 150 70 0504 47 73 0638 13 73 0504 095 73100 0724 135 75 0253 152 72 0503 52 73 0635 14 74 0503 096 75

100 25

25 0661 210 97 0355 240 97 0711 38 98 0611 11 95 0511 131 9850 0658 216 97 0355 244 99 0709 42 98 0606 15 97 0509 133 9875 0642 217 99 0333 247 100 0702 56 99 0602 18 98 0502 136 99100 0641 220 100 0331 250 100 0704 63 98 0602 21 98 0504 137 100

125 30

25 0754 487 120 0549 384 122 0829 27 122 0589 15 122 0529 167 12350 0748 486 121 0548 385 123 0824 45 123 0580 22 123 0524 168 12475 0750 489 122 0534 386 124 0822 59 125 0580 28 123 0522 170 125100 0752 495 125 0534 389 124 0822 65 124 0572 33 125 0522 172 125

150 35

25 0625 541 145 0766 588 147 0911 25 149 0559 28 148 0511 212 14950 0622 542 146 0765 561 148 0909 42 150 0547 36 149 0509 214 14975 0619 544 148 0763 564 149 0904 64 150 0523 43 150 0504 216 150100 0616 545 150 0763 569 149 0904 72 149 0523 48 150 0504 218 150

Table 3 Summary of experimental results of the different localization algorithms

Target nodes Anchor nodesPSO BOA FA GWO SSA

EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10 079 040 18 022 038 25 025 18 19 074 054 21 045 035 2450 15 043 076 46 032 086 49 042 25 49 069 081 46 046 069 4875 20 072 135 75 025 152 68 051 47 73 064 095 72 050 096 73100 25 065 216 100 035 245 100 070 53 98 060 21 98 051 135 100125 30 074 490 123 054 387 124 082 65 124 058 28 123 052 170 125150 35 062 543 149 076 565 149 090 72 149 052 43 150 050 215 150

Journal of Computer Networks and Communications 9

0

01

02

03

04M

ean

loca

lizat

ion

erro

r (m

)05

06

07

08

09

1

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

Figure 4 e localization error of the dierent localization algorithms in dierent WSN deployments

0

1

2

3

4

5

6

7

8

9

10

Com

putin

g tim

e (S)

(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35

Figure 5 e computing time of the dierent localization algorithms in dierent WSN deployments

0 20 40 60 80 100 120 140Number of localized nodes

(Num

ber o

f tar

get n

odes

ndashnu

mbe

r of

anch

ors)

SSAGWOFA

BOAPSO

25 ndash 10

50 ndash 15

75 ndash 20

100 ndash 25

125 ndash 30

150 ndash 35

Figure 6 e number of localized nodes of the dierent localization algorithms in dierent WSN deployments

10 Journal of Computer Networks and Communications

can be hybridized with other algorithm to reduce the local-ization error

Data Availability

No data were used to support this study Because our articlediscusses the problem of localization in WSNs our exper-imental results have been applied based on several networksrsquosize

Conflicts of Interest

+e authors declare that they have no conflicts of interest

References

[1] R V Kulkarni G K Venayagamoorthy and M X ChengldquoBio-inspired node localization in wireless sensor networksrdquoin Proceedings of IEEE International Conference in SystemsMan and Cybernetics (SMC) pp 205ndash210 San Antonio TXUSA October 2009

[2] D Lavanya and S K Udgata ldquoSwarm intelligence-basedlocalization in wireless sensor networksrdquo in InternationalWorkshop on Multi-Disciplinary Trends in Artificial In-telligence pp 317ndash328 Springer Berlin Germany 2011

[3] J Wang R K Ghosh and S K Das ldquoA survey on sensorlocalizationrdquo Journal of Control Aeory and Applicationsvol 8 no 1 pp 2ndash11 2010

[4] J Aspnes T Eren D K Goldenberg et al ldquoA theory ofnetwork localizationrdquo IEEE Transactions on Mobile Com-puting vol 5 no 12 pp 1663ndash1678 2006

[5] S Goyal and M S Patterh ldquoModified bat algorithm for lo-calization of wireless sensor networkrdquo Wireless PersonalCommunications vol 86 no 2 pp 657ndash670 2015

[6] M M Ahmed E H Houssein A E Hassanien A Taha andE Hassanien ldquoMaximizing lifetime of wireless sensor net-works based on whale optimization algorithmrdquo in In-ternational Conference on Advanced Intelligent Systems andInformatics pp 724ndash733 Springer Berlin Germany 2017

[7] K S Low H A Nguyen and H Guo ldquoA particle swarmoptimization approach for the localization of a wireless sensornetworkrdquo in Proceedings of IEEE International Symposium inIndustrial Electronics (ISIE) pp 1820ndash1825 VancouverCanada June 2008

[8] D Manjarres J Del Ser S Gil-Lopez M Vecchio I Landa-Torres and R Lopez-Valcarce ldquoOn the application of a hybridharmony search algorithm to node localization in anchor-based wireless sensor networksrdquo in Proceedings of IEEE In-ternational Conference in Intelligent Systems Design andApplications (ISDA) pp 1014ndash1019 Cordoba Spain No-vember 2011

[9] M Li W Xiong and Q Liang ldquoAn improved abc-based nodelocalization algorithm for wireless sensor networkrdquo in Pro-ceedings of IEEE International Conference in Wireless Com-munications Networking and Mobile Computing (WiCOM)pp 1ndash4 Barcelona Spain September 2012

[10] A Tamizharasi R Arthi and K Murugan ldquoBio-inspiredalgorithm for optimizing the localization of wireless sensornetworksrdquo in Proceedings of IEEE International Conference inComputing Communications and Networking Technologies(ICCCNT) pp 1ndash5 Tiruchengode India June 2013

[11] C Tang R Liu and J Ni ldquoA novel wireless sensor networklocalization approach localization based on plant growth

simulation algorithmrdquo Elektronika ir Elektrotechnika vol 19no 8 pp 97ndash100 2013

[12] O D Jegede and K Ferens ldquoA genetic algorithm for nodelocalization in wireless sensor networksrdquo in Proceedings ofInternational Conference on Genetic and EvolutionaryMethods (GEM) Ae Steering Committee of the World Con-gress in Computer Science Computer Engineering and AppliedComputing (WorldComp) pp 1ndash7 Las Vegas NV USA July2013

[13] S Goyal and M S Patterh ldquoWireless sensor network local-ization based on cuckoo search algorithmrdquo Wireless personalcommunications vol 79 no 1 pp 223ndash234 2014

[14] L Dan and W Xian-bin ldquoAn improved PSO algorithm fordistributed localization in wireless sensor networksrdquo In-ternational Journal of Distributed Sensor Networks vol 11no 7 Article ID 970272 2015

[15] R Krishnaprabha and A Gopakumar ldquoPerformance ofgravitational search algorithm in wireless sensor networklocalizationrdquo in Proceedings of IEEE National Conference inCommunication Signal Processing and Networking (NCCSN)pp 1ndash6 Palakkad India October 2014

[16] B Peng and L Li ldquoAn improved localization algorithm basedon genetic algorithm in wireless sensor networksrdquo CognitiveNeurodynamics vol 9 no 2 pp 249ndash256 2015

[17] V O Sai C S Shieh T T Nguyen Y C Lin M F Horngand Q D Le ldquoParallel firefly algorithm for localization al-gorithm in wireless sensor networkrdquo in Proceedings of IEEEInternational Conference in Robot Vision and Signal Pro-cessing (RVSP) pp 300ndash305 Kaohsiung China November2015

[18] S Sivakumar and R Venkatesan ldquoMeta-heuristic approachesfor minimizing error in localization of wireless sensor net-worksrdquo Applied Soft Computing vol 36 pp 506ndash518 2015

[19] Z Sun L Tao X Wang and Z Zhou ldquoLocalization algorithmin wireless sensor networks based on multiobjective particleswarm optimizationrdquo International Journal of DistributedSensor Networks vol 11 no 8 Article ID 716291 2015

[20] A Arsic M Tuba and M Jordanski ldquoFireworks algorithmapplied to wireless sensor networks localization problemrdquo inProceedings of IEEE Congress on Evolutionary Computation(CEC) pp 4038ndash4044 Vancouver Canada July 2016

[21] C S Shieh V O Sai Y C Lin T F Lee T T Nguyen andQ D Le ldquoImproved node localization for WSN using heu-ristic optimization approachesrdquo in Proceedings of IEEE In-ternational Conference in Networking and NetworkApplications (NaNA) pp 95ndash98 Hokkaido Japan July 2016

[22] J Cheng and L Xia ldquoAn effective Cuckoo search algorithm fornode localization in wireless sensor networkrdquo Sensors vol 16no 9 p 1390 2016

[23] P T Daely and S Y Shin ldquoRange based wireless node lo-calization using Dragonfly Algorithmrdquo in Proceedings of IEEEInternational Conference in Ubiquitous and Future Networks(ICUFN) pp 1012ndash1015 Vienna Austria July 2016

[24] T T Nguyen J S Pan S C Chu J F Roddick and T K DaoldquoOptimization localization in wireless sensor network basedon multi-objective firefly algorithmrdquo Journal of NetworkIntelligence vol 1 no 4 pp 130ndash138 2016

[25] S Arora and S Singh ldquoNode localization in wireless sensornetworks using butterfly optimization algorithmrdquo ArabianJournal for Science and Engineering vol 42 no 8pp 3325ndash3335 2017

[26] S R Sujatha and M Siddappa ldquoNode localization method forwireless sensor networks based on hybrid optimization ofparticle swarm optimization and differential evolutionrdquo IOSR

Journal of Computer Networks and Communications 11

Journal of Computer Engineering vol 19 no 2 pp 07ndash122017

[27] R Kaur and S Arora ldquonature inspired range based wirelesssensor node localization algorithmsrdquo International Journal ofInteractive Multimedia and Artificial Intelligence vol 4pp 7ndash17 2017

[28] A +arwat E H Houssein M M Ahmed A E Hassanienand T Gabel ldquoMOGOA algorithm for constrained and un-constrained multi-objective optimization problemsrdquo AppliedIntelligence vol 48 no 8 pp 2268ndash2283 2017

[29] A A Ewees M Abd Elaziz and E H Houssein ldquoImprovedgrasshopper optimization algorithm using opposition-basedlearningrdquo Expert Systems with Applications vol 112pp 156ndash172 2018

[30] A G Hussien E H Houssein and A E Hassanien ldquoA binarywhale optimization algorithm with hyperbolic tangent fitnessfunction for feature selectionrdquo in Proceedings of Eighth In-ternational Conference on Intelligent Computing and In-formation Systems (ICICIS) pp 166ndash172 Cairo EgyptDecember 2017

[31] B Xue M Zhang and W N Browne ldquoParticle swarm op-timization for feature selection in classification a multi-objective approachrdquo IEEE Transactions on Cyberneticsvol 43 no 6 pp 1656ndash1671 2013

[32] S Arora and S Singh ldquoButterfly algorithmwith levy flights forglobal optimizationrdquo in Proceedings of IEEE InternationalConference in Signal Processing Computing and Control(ISPCC) pp 220ndash224 Waknaghat India September 2015

[33] S Arora and S Singh ldquoAn improved butterfly optimizationalgorithm with chaosrdquo Journal of Intelligent and Fuzzy Sys-tems vol 32 no 1 pp 1079ndash1088 2017

[34] R B Blair and A E Launer ldquoButterfly diversity and humanland use species assemblages along an urban grandientrdquoBiological Conservation vol 80 no 1 pp 113ndash125 1997

[35] X S Yang and X He ldquoFirefly algorithm recent advances andapplicationsrdquo 2013 httpsarxivorgabs13083898

[36] X S Yang ldquoFirefly algorithms for multimodal optimizationrdquoin Proceedings of International Symposium on Stochastic Al-gorithms pp 169ndash178 Sapporo Japan October 2009

[37] S Mirjalili S M Mirjalili and A Lewis ldquoGrey wolf opti-mizerrdquo Advances in Engineering Software vol 69 pp 46ndash612014

[38] L P Madin ldquoAspects of jet propulsion in Salpsrdquo CanadianJournal of Zoology vol 68 no 4 pp 765ndash777 1990

[39] S Mirjalili A H Gandomi S Z Mirjalili S Saremi H Farisand S M Mirjalili ldquoSalp swarm algorithm a bio-inspiredoptimizer for engineering design problemsrdquo Advances inEngineering Software vol 114 pp 163ndash191 2017

[40] P A Anderson and Q Bone ldquoCommunication betweenindividuals in Salp chains II Physiologyrdquo Proceedings of theRoyal Society of London Series B Biological Sciences vol 210no 1181 pp 559ndash574 1980

[41] H T Ibrahim W J Mazher O N Ucan and O BayatldquoFeature selection using salp swarm algorithm for real bio-medical datasetsrdquo IJCSNS vol 17 no 12 pp 13ndash20 2017

12 Journal of Computer Networks and Communications

nodes are increased Regarding the computing time it isnoticed that increasing the number of target nodes andanchor nodes increase the computing time for all local-ization algorithms However once again SSA has the bestcomputing time compared to PSO BOA FA and GWOFinally regarding the number of localized nodes (NL) it isnoticed that SSA has the best results compared to otherlocalization algorithms In addition it is noticed that SSAhas inferior EL when the percentage of (number of anchornodesnumber of target nodes) becomes larger such as(1025) in the first case in Table 3 +at is because thelocalization accuracy increases when the anchor densityincreases with respect to the number of the target nodes[25]

+e graphical representations of the experimentalresults for the different performance metrics are shown inFigures 4ndash6

6 Conclusion

Accurate node localization concerns many applications thatadopt WSNs In this paper a node localization algorithm hasbeen proposed based on a novel bioinspired algorithm calledSalp Swarm Algorithm (SSA) which handled the node lo-calization problem as an optimization problem+e proposedalgorithm has been implemented and validated in differentWSN deployments using different numbers of target nodesand anchor nodes Moreover the proposed algorithm hasbeen evaluated and compared to four well-known optimi-zation algorithms namely PSO BOA FA and GWO interms of localization accuracy computing time and severallocalized nodes +e obtained simulation results have provedthe superiority of the proposed algorithm compared to theother localization algorithms regarding the different perfor-mance metrics In the future work the proposed approach

Table 2 Performance metrics of different localization algorithms

Targetnodes

Anchornodes

No ofiterations

PSO BOA FA GWO SSAEL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10

25 0818 040 16 0232 037 22 0265 06 19 0744 022 20 0465 035 2250 0812 040 15 0225 038 23 0262 12 20 0741 041 21 0462 035 2375 0803 041 18 0223 039 25 0258 15 21 0741 054 21 0458 036 23100 0792 041 18 0221 040 25 0251 18 19 0740 079 23 0451 037 24

50 15

25 0419 071 41 0338 084 47 0477 16 46 0690 042 44 0477 067 4350 0426 073 47 0332 085 48 0473 19 49 0688 063 45 0472 069 4775 0429 076 46 0326 086 49 0465 25 49 0686 081 46 0468 069 48100 0434 076 48 0323 086 49 0465 35 48 0682 098 48 0464 070 50

75 20

25 0735 131 73 0257 149 66 0519 29 71 0641 072 72 0519 090 6950 0728 132 74 0257 149 66 0513 38 72 0641 095 72 0513 092 7275 0728 133 74 0255 150 70 0504 47 73 0638 13 73 0504 095 73100 0724 135 75 0253 152 72 0503 52 73 0635 14 74 0503 096 75

100 25

25 0661 210 97 0355 240 97 0711 38 98 0611 11 95 0511 131 9850 0658 216 97 0355 244 99 0709 42 98 0606 15 97 0509 133 9875 0642 217 99 0333 247 100 0702 56 99 0602 18 98 0502 136 99100 0641 220 100 0331 250 100 0704 63 98 0602 21 98 0504 137 100

125 30

25 0754 487 120 0549 384 122 0829 27 122 0589 15 122 0529 167 12350 0748 486 121 0548 385 123 0824 45 123 0580 22 123 0524 168 12475 0750 489 122 0534 386 124 0822 59 125 0580 28 123 0522 170 125100 0752 495 125 0534 389 124 0822 65 124 0572 33 125 0522 172 125

150 35

25 0625 541 145 0766 588 147 0911 25 149 0559 28 148 0511 212 14950 0622 542 146 0765 561 148 0909 42 150 0547 36 149 0509 214 14975 0619 544 148 0763 564 149 0904 64 150 0523 43 150 0504 216 150100 0616 545 150 0763 569 149 0904 72 149 0523 48 150 0504 218 150

Table 3 Summary of experimental results of the different localization algorithms

Target nodes Anchor nodesPSO BOA FA GWO SSA

EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL EL(m) T(s) NL

25 10 079 040 18 022 038 25 025 18 19 074 054 21 045 035 2450 15 043 076 46 032 086 49 042 25 49 069 081 46 046 069 4875 20 072 135 75 025 152 68 051 47 73 064 095 72 050 096 73100 25 065 216 100 035 245 100 070 53 98 060 21 98 051 135 100125 30 074 490 123 054 387 124 082 65 124 058 28 123 052 170 125150 35 062 543 149 076 565 149 090 72 149 052 43 150 050 215 150

Journal of Computer Networks and Communications 9

0

01

02

03

04M

ean

loca

lizat

ion

erro

r (m

)05

06

07

08

09

1

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

Figure 4 e localization error of the dierent localization algorithms in dierent WSN deployments

0

1

2

3

4

5

6

7

8

9

10

Com

putin

g tim

e (S)

(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35

Figure 5 e computing time of the dierent localization algorithms in dierent WSN deployments

0 20 40 60 80 100 120 140Number of localized nodes

(Num

ber o

f tar

get n

odes

ndashnu

mbe

r of

anch

ors)

SSAGWOFA

BOAPSO

25 ndash 10

50 ndash 15

75 ndash 20

100 ndash 25

125 ndash 30

150 ndash 35

Figure 6 e number of localized nodes of the dierent localization algorithms in dierent WSN deployments

10 Journal of Computer Networks and Communications

can be hybridized with other algorithm to reduce the local-ization error

Data Availability

No data were used to support this study Because our articlediscusses the problem of localization in WSNs our exper-imental results have been applied based on several networksrsquosize

Conflicts of Interest

+e authors declare that they have no conflicts of interest

References

[1] R V Kulkarni G K Venayagamoorthy and M X ChengldquoBio-inspired node localization in wireless sensor networksrdquoin Proceedings of IEEE International Conference in SystemsMan and Cybernetics (SMC) pp 205ndash210 San Antonio TXUSA October 2009

[2] D Lavanya and S K Udgata ldquoSwarm intelligence-basedlocalization in wireless sensor networksrdquo in InternationalWorkshop on Multi-Disciplinary Trends in Artificial In-telligence pp 317ndash328 Springer Berlin Germany 2011

[3] J Wang R K Ghosh and S K Das ldquoA survey on sensorlocalizationrdquo Journal of Control Aeory and Applicationsvol 8 no 1 pp 2ndash11 2010

[4] J Aspnes T Eren D K Goldenberg et al ldquoA theory ofnetwork localizationrdquo IEEE Transactions on Mobile Com-puting vol 5 no 12 pp 1663ndash1678 2006

[5] S Goyal and M S Patterh ldquoModified bat algorithm for lo-calization of wireless sensor networkrdquo Wireless PersonalCommunications vol 86 no 2 pp 657ndash670 2015

[6] M M Ahmed E H Houssein A E Hassanien A Taha andE Hassanien ldquoMaximizing lifetime of wireless sensor net-works based on whale optimization algorithmrdquo in In-ternational Conference on Advanced Intelligent Systems andInformatics pp 724ndash733 Springer Berlin Germany 2017

[7] K S Low H A Nguyen and H Guo ldquoA particle swarmoptimization approach for the localization of a wireless sensornetworkrdquo in Proceedings of IEEE International Symposium inIndustrial Electronics (ISIE) pp 1820ndash1825 VancouverCanada June 2008

[8] D Manjarres J Del Ser S Gil-Lopez M Vecchio I Landa-Torres and R Lopez-Valcarce ldquoOn the application of a hybridharmony search algorithm to node localization in anchor-based wireless sensor networksrdquo in Proceedings of IEEE In-ternational Conference in Intelligent Systems Design andApplications (ISDA) pp 1014ndash1019 Cordoba Spain No-vember 2011

[9] M Li W Xiong and Q Liang ldquoAn improved abc-based nodelocalization algorithm for wireless sensor networkrdquo in Pro-ceedings of IEEE International Conference in Wireless Com-munications Networking and Mobile Computing (WiCOM)pp 1ndash4 Barcelona Spain September 2012

[10] A Tamizharasi R Arthi and K Murugan ldquoBio-inspiredalgorithm for optimizing the localization of wireless sensornetworksrdquo in Proceedings of IEEE International Conference inComputing Communications and Networking Technologies(ICCCNT) pp 1ndash5 Tiruchengode India June 2013

[11] C Tang R Liu and J Ni ldquoA novel wireless sensor networklocalization approach localization based on plant growth

simulation algorithmrdquo Elektronika ir Elektrotechnika vol 19no 8 pp 97ndash100 2013

[12] O D Jegede and K Ferens ldquoA genetic algorithm for nodelocalization in wireless sensor networksrdquo in Proceedings ofInternational Conference on Genetic and EvolutionaryMethods (GEM) Ae Steering Committee of the World Con-gress in Computer Science Computer Engineering and AppliedComputing (WorldComp) pp 1ndash7 Las Vegas NV USA July2013

[13] S Goyal and M S Patterh ldquoWireless sensor network local-ization based on cuckoo search algorithmrdquo Wireless personalcommunications vol 79 no 1 pp 223ndash234 2014

[14] L Dan and W Xian-bin ldquoAn improved PSO algorithm fordistributed localization in wireless sensor networksrdquo In-ternational Journal of Distributed Sensor Networks vol 11no 7 Article ID 970272 2015

[15] R Krishnaprabha and A Gopakumar ldquoPerformance ofgravitational search algorithm in wireless sensor networklocalizationrdquo in Proceedings of IEEE National Conference inCommunication Signal Processing and Networking (NCCSN)pp 1ndash6 Palakkad India October 2014

[16] B Peng and L Li ldquoAn improved localization algorithm basedon genetic algorithm in wireless sensor networksrdquo CognitiveNeurodynamics vol 9 no 2 pp 249ndash256 2015

[17] V O Sai C S Shieh T T Nguyen Y C Lin M F Horngand Q D Le ldquoParallel firefly algorithm for localization al-gorithm in wireless sensor networkrdquo in Proceedings of IEEEInternational Conference in Robot Vision and Signal Pro-cessing (RVSP) pp 300ndash305 Kaohsiung China November2015

[18] S Sivakumar and R Venkatesan ldquoMeta-heuristic approachesfor minimizing error in localization of wireless sensor net-worksrdquo Applied Soft Computing vol 36 pp 506ndash518 2015

[19] Z Sun L Tao X Wang and Z Zhou ldquoLocalization algorithmin wireless sensor networks based on multiobjective particleswarm optimizationrdquo International Journal of DistributedSensor Networks vol 11 no 8 Article ID 716291 2015

[20] A Arsic M Tuba and M Jordanski ldquoFireworks algorithmapplied to wireless sensor networks localization problemrdquo inProceedings of IEEE Congress on Evolutionary Computation(CEC) pp 4038ndash4044 Vancouver Canada July 2016

[21] C S Shieh V O Sai Y C Lin T F Lee T T Nguyen andQ D Le ldquoImproved node localization for WSN using heu-ristic optimization approachesrdquo in Proceedings of IEEE In-ternational Conference in Networking and NetworkApplications (NaNA) pp 95ndash98 Hokkaido Japan July 2016

[22] J Cheng and L Xia ldquoAn effective Cuckoo search algorithm fornode localization in wireless sensor networkrdquo Sensors vol 16no 9 p 1390 2016

[23] P T Daely and S Y Shin ldquoRange based wireless node lo-calization using Dragonfly Algorithmrdquo in Proceedings of IEEEInternational Conference in Ubiquitous and Future Networks(ICUFN) pp 1012ndash1015 Vienna Austria July 2016

[24] T T Nguyen J S Pan S C Chu J F Roddick and T K DaoldquoOptimization localization in wireless sensor network basedon multi-objective firefly algorithmrdquo Journal of NetworkIntelligence vol 1 no 4 pp 130ndash138 2016

[25] S Arora and S Singh ldquoNode localization in wireless sensornetworks using butterfly optimization algorithmrdquo ArabianJournal for Science and Engineering vol 42 no 8pp 3325ndash3335 2017

[26] S R Sujatha and M Siddappa ldquoNode localization method forwireless sensor networks based on hybrid optimization ofparticle swarm optimization and differential evolutionrdquo IOSR

Journal of Computer Networks and Communications 11

Journal of Computer Engineering vol 19 no 2 pp 07ndash122017

[27] R Kaur and S Arora ldquonature inspired range based wirelesssensor node localization algorithmsrdquo International Journal ofInteractive Multimedia and Artificial Intelligence vol 4pp 7ndash17 2017

[28] A +arwat E H Houssein M M Ahmed A E Hassanienand T Gabel ldquoMOGOA algorithm for constrained and un-constrained multi-objective optimization problemsrdquo AppliedIntelligence vol 48 no 8 pp 2268ndash2283 2017

[29] A A Ewees M Abd Elaziz and E H Houssein ldquoImprovedgrasshopper optimization algorithm using opposition-basedlearningrdquo Expert Systems with Applications vol 112pp 156ndash172 2018

[30] A G Hussien E H Houssein and A E Hassanien ldquoA binarywhale optimization algorithm with hyperbolic tangent fitnessfunction for feature selectionrdquo in Proceedings of Eighth In-ternational Conference on Intelligent Computing and In-formation Systems (ICICIS) pp 166ndash172 Cairo EgyptDecember 2017

[31] B Xue M Zhang and W N Browne ldquoParticle swarm op-timization for feature selection in classification a multi-objective approachrdquo IEEE Transactions on Cyberneticsvol 43 no 6 pp 1656ndash1671 2013

[32] S Arora and S Singh ldquoButterfly algorithmwith levy flights forglobal optimizationrdquo in Proceedings of IEEE InternationalConference in Signal Processing Computing and Control(ISPCC) pp 220ndash224 Waknaghat India September 2015

[33] S Arora and S Singh ldquoAn improved butterfly optimizationalgorithm with chaosrdquo Journal of Intelligent and Fuzzy Sys-tems vol 32 no 1 pp 1079ndash1088 2017

[34] R B Blair and A E Launer ldquoButterfly diversity and humanland use species assemblages along an urban grandientrdquoBiological Conservation vol 80 no 1 pp 113ndash125 1997

[35] X S Yang and X He ldquoFirefly algorithm recent advances andapplicationsrdquo 2013 httpsarxivorgabs13083898

[36] X S Yang ldquoFirefly algorithms for multimodal optimizationrdquoin Proceedings of International Symposium on Stochastic Al-gorithms pp 169ndash178 Sapporo Japan October 2009

[37] S Mirjalili S M Mirjalili and A Lewis ldquoGrey wolf opti-mizerrdquo Advances in Engineering Software vol 69 pp 46ndash612014

[38] L P Madin ldquoAspects of jet propulsion in Salpsrdquo CanadianJournal of Zoology vol 68 no 4 pp 765ndash777 1990

[39] S Mirjalili A H Gandomi S Z Mirjalili S Saremi H Farisand S M Mirjalili ldquoSalp swarm algorithm a bio-inspiredoptimizer for engineering design problemsrdquo Advances inEngineering Software vol 114 pp 163ndash191 2017

[40] P A Anderson and Q Bone ldquoCommunication betweenindividuals in Salp chains II Physiologyrdquo Proceedings of theRoyal Society of London Series B Biological Sciences vol 210no 1181 pp 559ndash574 1980

[41] H T Ibrahim W J Mazher O N Ucan and O BayatldquoFeature selection using salp swarm algorithm for real bio-medical datasetsrdquo IJCSNS vol 17 no 12 pp 13ndash20 2017

12 Journal of Computer Networks and Communications

0

01

02

03

04M

ean

loca

lizat

ion

erro

r (m

)05

06

07

08

09

1

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

Figure 4 e localization error of the dierent localization algorithms in dierent WSN deployments

0

1

2

3

4

5

6

7

8

9

10

Com

putin

g tim

e (S)

(Number of target nodes ndash number of anchors)

PSOBOAFA

GWOSSA

25 ndash 10 50 ndash 15 75 ndash 20 100 ndash 25 125 ndash 30 150 ndash 35

Figure 5 e computing time of the dierent localization algorithms in dierent WSN deployments

0 20 40 60 80 100 120 140Number of localized nodes

(Num

ber o

f tar

get n

odes

ndashnu

mbe

r of

anch

ors)

SSAGWOFA

BOAPSO

25 ndash 10

50 ndash 15

75 ndash 20

100 ndash 25

125 ndash 30

150 ndash 35

Figure 6 e number of localized nodes of the dierent localization algorithms in dierent WSN deployments

10 Journal of Computer Networks and Communications

can be hybridized with other algorithm to reduce the local-ization error

Data Availability

No data were used to support this study Because our articlediscusses the problem of localization in WSNs our exper-imental results have been applied based on several networksrsquosize

Conflicts of Interest

+e authors declare that they have no conflicts of interest

References

[1] R V Kulkarni G K Venayagamoorthy and M X ChengldquoBio-inspired node localization in wireless sensor networksrdquoin Proceedings of IEEE International Conference in SystemsMan and Cybernetics (SMC) pp 205ndash210 San Antonio TXUSA October 2009

[2] D Lavanya and S K Udgata ldquoSwarm intelligence-basedlocalization in wireless sensor networksrdquo in InternationalWorkshop on Multi-Disciplinary Trends in Artificial In-telligence pp 317ndash328 Springer Berlin Germany 2011

[3] J Wang R K Ghosh and S K Das ldquoA survey on sensorlocalizationrdquo Journal of Control Aeory and Applicationsvol 8 no 1 pp 2ndash11 2010

[4] J Aspnes T Eren D K Goldenberg et al ldquoA theory ofnetwork localizationrdquo IEEE Transactions on Mobile Com-puting vol 5 no 12 pp 1663ndash1678 2006

[5] S Goyal and M S Patterh ldquoModified bat algorithm for lo-calization of wireless sensor networkrdquo Wireless PersonalCommunications vol 86 no 2 pp 657ndash670 2015

[6] M M Ahmed E H Houssein A E Hassanien A Taha andE Hassanien ldquoMaximizing lifetime of wireless sensor net-works based on whale optimization algorithmrdquo in In-ternational Conference on Advanced Intelligent Systems andInformatics pp 724ndash733 Springer Berlin Germany 2017

[7] K S Low H A Nguyen and H Guo ldquoA particle swarmoptimization approach for the localization of a wireless sensornetworkrdquo in Proceedings of IEEE International Symposium inIndustrial Electronics (ISIE) pp 1820ndash1825 VancouverCanada June 2008

[8] D Manjarres J Del Ser S Gil-Lopez M Vecchio I Landa-Torres and R Lopez-Valcarce ldquoOn the application of a hybridharmony search algorithm to node localization in anchor-based wireless sensor networksrdquo in Proceedings of IEEE In-ternational Conference in Intelligent Systems Design andApplications (ISDA) pp 1014ndash1019 Cordoba Spain No-vember 2011

[9] M Li W Xiong and Q Liang ldquoAn improved abc-based nodelocalization algorithm for wireless sensor networkrdquo in Pro-ceedings of IEEE International Conference in Wireless Com-munications Networking and Mobile Computing (WiCOM)pp 1ndash4 Barcelona Spain September 2012

[10] A Tamizharasi R Arthi and K Murugan ldquoBio-inspiredalgorithm for optimizing the localization of wireless sensornetworksrdquo in Proceedings of IEEE International Conference inComputing Communications and Networking Technologies(ICCCNT) pp 1ndash5 Tiruchengode India June 2013

[11] C Tang R Liu and J Ni ldquoA novel wireless sensor networklocalization approach localization based on plant growth

simulation algorithmrdquo Elektronika ir Elektrotechnika vol 19no 8 pp 97ndash100 2013

[12] O D Jegede and K Ferens ldquoA genetic algorithm for nodelocalization in wireless sensor networksrdquo in Proceedings ofInternational Conference on Genetic and EvolutionaryMethods (GEM) Ae Steering Committee of the World Con-gress in Computer Science Computer Engineering and AppliedComputing (WorldComp) pp 1ndash7 Las Vegas NV USA July2013

[13] S Goyal and M S Patterh ldquoWireless sensor network local-ization based on cuckoo search algorithmrdquo Wireless personalcommunications vol 79 no 1 pp 223ndash234 2014

[14] L Dan and W Xian-bin ldquoAn improved PSO algorithm fordistributed localization in wireless sensor networksrdquo In-ternational Journal of Distributed Sensor Networks vol 11no 7 Article ID 970272 2015

[15] R Krishnaprabha and A Gopakumar ldquoPerformance ofgravitational search algorithm in wireless sensor networklocalizationrdquo in Proceedings of IEEE National Conference inCommunication Signal Processing and Networking (NCCSN)pp 1ndash6 Palakkad India October 2014

[16] B Peng and L Li ldquoAn improved localization algorithm basedon genetic algorithm in wireless sensor networksrdquo CognitiveNeurodynamics vol 9 no 2 pp 249ndash256 2015

[17] V O Sai C S Shieh T T Nguyen Y C Lin M F Horngand Q D Le ldquoParallel firefly algorithm for localization al-gorithm in wireless sensor networkrdquo in Proceedings of IEEEInternational Conference in Robot Vision and Signal Pro-cessing (RVSP) pp 300ndash305 Kaohsiung China November2015

[18] S Sivakumar and R Venkatesan ldquoMeta-heuristic approachesfor minimizing error in localization of wireless sensor net-worksrdquo Applied Soft Computing vol 36 pp 506ndash518 2015

[19] Z Sun L Tao X Wang and Z Zhou ldquoLocalization algorithmin wireless sensor networks based on multiobjective particleswarm optimizationrdquo International Journal of DistributedSensor Networks vol 11 no 8 Article ID 716291 2015

[20] A Arsic M Tuba and M Jordanski ldquoFireworks algorithmapplied to wireless sensor networks localization problemrdquo inProceedings of IEEE Congress on Evolutionary Computation(CEC) pp 4038ndash4044 Vancouver Canada July 2016

[21] C S Shieh V O Sai Y C Lin T F Lee T T Nguyen andQ D Le ldquoImproved node localization for WSN using heu-ristic optimization approachesrdquo in Proceedings of IEEE In-ternational Conference in Networking and NetworkApplications (NaNA) pp 95ndash98 Hokkaido Japan July 2016

[22] J Cheng and L Xia ldquoAn effective Cuckoo search algorithm fornode localization in wireless sensor networkrdquo Sensors vol 16no 9 p 1390 2016

[23] P T Daely and S Y Shin ldquoRange based wireless node lo-calization using Dragonfly Algorithmrdquo in Proceedings of IEEEInternational Conference in Ubiquitous and Future Networks(ICUFN) pp 1012ndash1015 Vienna Austria July 2016

[24] T T Nguyen J S Pan S C Chu J F Roddick and T K DaoldquoOptimization localization in wireless sensor network basedon multi-objective firefly algorithmrdquo Journal of NetworkIntelligence vol 1 no 4 pp 130ndash138 2016

[25] S Arora and S Singh ldquoNode localization in wireless sensornetworks using butterfly optimization algorithmrdquo ArabianJournal for Science and Engineering vol 42 no 8pp 3325ndash3335 2017

[26] S R Sujatha and M Siddappa ldquoNode localization method forwireless sensor networks based on hybrid optimization ofparticle swarm optimization and differential evolutionrdquo IOSR

Journal of Computer Networks and Communications 11

Journal of Computer Engineering vol 19 no 2 pp 07ndash122017

[27] R Kaur and S Arora ldquonature inspired range based wirelesssensor node localization algorithmsrdquo International Journal ofInteractive Multimedia and Artificial Intelligence vol 4pp 7ndash17 2017

[28] A +arwat E H Houssein M M Ahmed A E Hassanienand T Gabel ldquoMOGOA algorithm for constrained and un-constrained multi-objective optimization problemsrdquo AppliedIntelligence vol 48 no 8 pp 2268ndash2283 2017

[29] A A Ewees M Abd Elaziz and E H Houssein ldquoImprovedgrasshopper optimization algorithm using opposition-basedlearningrdquo Expert Systems with Applications vol 112pp 156ndash172 2018

[30] A G Hussien E H Houssein and A E Hassanien ldquoA binarywhale optimization algorithm with hyperbolic tangent fitnessfunction for feature selectionrdquo in Proceedings of Eighth In-ternational Conference on Intelligent Computing and In-formation Systems (ICICIS) pp 166ndash172 Cairo EgyptDecember 2017

[31] B Xue M Zhang and W N Browne ldquoParticle swarm op-timization for feature selection in classification a multi-objective approachrdquo IEEE Transactions on Cyberneticsvol 43 no 6 pp 1656ndash1671 2013

[32] S Arora and S Singh ldquoButterfly algorithmwith levy flights forglobal optimizationrdquo in Proceedings of IEEE InternationalConference in Signal Processing Computing and Control(ISPCC) pp 220ndash224 Waknaghat India September 2015

[33] S Arora and S Singh ldquoAn improved butterfly optimizationalgorithm with chaosrdquo Journal of Intelligent and Fuzzy Sys-tems vol 32 no 1 pp 1079ndash1088 2017

[34] R B Blair and A E Launer ldquoButterfly diversity and humanland use species assemblages along an urban grandientrdquoBiological Conservation vol 80 no 1 pp 113ndash125 1997

[35] X S Yang and X He ldquoFirefly algorithm recent advances andapplicationsrdquo 2013 httpsarxivorgabs13083898

[36] X S Yang ldquoFirefly algorithms for multimodal optimizationrdquoin Proceedings of International Symposium on Stochastic Al-gorithms pp 169ndash178 Sapporo Japan October 2009

[37] S Mirjalili S M Mirjalili and A Lewis ldquoGrey wolf opti-mizerrdquo Advances in Engineering Software vol 69 pp 46ndash612014

[38] L P Madin ldquoAspects of jet propulsion in Salpsrdquo CanadianJournal of Zoology vol 68 no 4 pp 765ndash777 1990

[39] S Mirjalili A H Gandomi S Z Mirjalili S Saremi H Farisand S M Mirjalili ldquoSalp swarm algorithm a bio-inspiredoptimizer for engineering design problemsrdquo Advances inEngineering Software vol 114 pp 163ndash191 2017

[40] P A Anderson and Q Bone ldquoCommunication betweenindividuals in Salp chains II Physiologyrdquo Proceedings of theRoyal Society of London Series B Biological Sciences vol 210no 1181 pp 559ndash574 1980

[41] H T Ibrahim W J Mazher O N Ucan and O BayatldquoFeature selection using salp swarm algorithm for real bio-medical datasetsrdquo IJCSNS vol 17 no 12 pp 13ndash20 2017

12 Journal of Computer Networks and Communications

can be hybridized with other algorithm to reduce the local-ization error

Data Availability

No data were used to support this study Because our articlediscusses the problem of localization in WSNs our exper-imental results have been applied based on several networksrsquosize

Conflicts of Interest

+e authors declare that they have no conflicts of interest

References

[1] R V Kulkarni G K Venayagamoorthy and M X ChengldquoBio-inspired node localization in wireless sensor networksrdquoin Proceedings of IEEE International Conference in SystemsMan and Cybernetics (SMC) pp 205ndash210 San Antonio TXUSA October 2009

[2] D Lavanya and S K Udgata ldquoSwarm intelligence-basedlocalization in wireless sensor networksrdquo in InternationalWorkshop on Multi-Disciplinary Trends in Artificial In-telligence pp 317ndash328 Springer Berlin Germany 2011

[3] J Wang R K Ghosh and S K Das ldquoA survey on sensorlocalizationrdquo Journal of Control Aeory and Applicationsvol 8 no 1 pp 2ndash11 2010

[4] J Aspnes T Eren D K Goldenberg et al ldquoA theory ofnetwork localizationrdquo IEEE Transactions on Mobile Com-puting vol 5 no 12 pp 1663ndash1678 2006

[5] S Goyal and M S Patterh ldquoModified bat algorithm for lo-calization of wireless sensor networkrdquo Wireless PersonalCommunications vol 86 no 2 pp 657ndash670 2015

[6] M M Ahmed E H Houssein A E Hassanien A Taha andE Hassanien ldquoMaximizing lifetime of wireless sensor net-works based on whale optimization algorithmrdquo in In-ternational Conference on Advanced Intelligent Systems andInformatics pp 724ndash733 Springer Berlin Germany 2017

[7] K S Low H A Nguyen and H Guo ldquoA particle swarmoptimization approach for the localization of a wireless sensornetworkrdquo in Proceedings of IEEE International Symposium inIndustrial Electronics (ISIE) pp 1820ndash1825 VancouverCanada June 2008

[8] D Manjarres J Del Ser S Gil-Lopez M Vecchio I Landa-Torres and R Lopez-Valcarce ldquoOn the application of a hybridharmony search algorithm to node localization in anchor-based wireless sensor networksrdquo in Proceedings of IEEE In-ternational Conference in Intelligent Systems Design andApplications (ISDA) pp 1014ndash1019 Cordoba Spain No-vember 2011

[9] M Li W Xiong and Q Liang ldquoAn improved abc-based nodelocalization algorithm for wireless sensor networkrdquo in Pro-ceedings of IEEE International Conference in Wireless Com-munications Networking and Mobile Computing (WiCOM)pp 1ndash4 Barcelona Spain September 2012

[10] A Tamizharasi R Arthi and K Murugan ldquoBio-inspiredalgorithm for optimizing the localization of wireless sensornetworksrdquo in Proceedings of IEEE International Conference inComputing Communications and Networking Technologies(ICCCNT) pp 1ndash5 Tiruchengode India June 2013

[11] C Tang R Liu and J Ni ldquoA novel wireless sensor networklocalization approach localization based on plant growth

simulation algorithmrdquo Elektronika ir Elektrotechnika vol 19no 8 pp 97ndash100 2013

[12] O D Jegede and K Ferens ldquoA genetic algorithm for nodelocalization in wireless sensor networksrdquo in Proceedings ofInternational Conference on Genetic and EvolutionaryMethods (GEM) Ae Steering Committee of the World Con-gress in Computer Science Computer Engineering and AppliedComputing (WorldComp) pp 1ndash7 Las Vegas NV USA July2013

[13] S Goyal and M S Patterh ldquoWireless sensor network local-ization based on cuckoo search algorithmrdquo Wireless personalcommunications vol 79 no 1 pp 223ndash234 2014

[14] L Dan and W Xian-bin ldquoAn improved PSO algorithm fordistributed localization in wireless sensor networksrdquo In-ternational Journal of Distributed Sensor Networks vol 11no 7 Article ID 970272 2015

[15] R Krishnaprabha and A Gopakumar ldquoPerformance ofgravitational search algorithm in wireless sensor networklocalizationrdquo in Proceedings of IEEE National Conference inCommunication Signal Processing and Networking (NCCSN)pp 1ndash6 Palakkad India October 2014

[16] B Peng and L Li ldquoAn improved localization algorithm basedon genetic algorithm in wireless sensor networksrdquo CognitiveNeurodynamics vol 9 no 2 pp 249ndash256 2015

[17] V O Sai C S Shieh T T Nguyen Y C Lin M F Horngand Q D Le ldquoParallel firefly algorithm for localization al-gorithm in wireless sensor networkrdquo in Proceedings of IEEEInternational Conference in Robot Vision and Signal Pro-cessing (RVSP) pp 300ndash305 Kaohsiung China November2015

[18] S Sivakumar and R Venkatesan ldquoMeta-heuristic approachesfor minimizing error in localization of wireless sensor net-worksrdquo Applied Soft Computing vol 36 pp 506ndash518 2015

[19] Z Sun L Tao X Wang and Z Zhou ldquoLocalization algorithmin wireless sensor networks based on multiobjective particleswarm optimizationrdquo International Journal of DistributedSensor Networks vol 11 no 8 Article ID 716291 2015

[20] A Arsic M Tuba and M Jordanski ldquoFireworks algorithmapplied to wireless sensor networks localization problemrdquo inProceedings of IEEE Congress on Evolutionary Computation(CEC) pp 4038ndash4044 Vancouver Canada July 2016

[21] C S Shieh V O Sai Y C Lin T F Lee T T Nguyen andQ D Le ldquoImproved node localization for WSN using heu-ristic optimization approachesrdquo in Proceedings of IEEE In-ternational Conference in Networking and NetworkApplications (NaNA) pp 95ndash98 Hokkaido Japan July 2016

[22] J Cheng and L Xia ldquoAn effective Cuckoo search algorithm fornode localization in wireless sensor networkrdquo Sensors vol 16no 9 p 1390 2016

[23] P T Daely and S Y Shin ldquoRange based wireless node lo-calization using Dragonfly Algorithmrdquo in Proceedings of IEEEInternational Conference in Ubiquitous and Future Networks(ICUFN) pp 1012ndash1015 Vienna Austria July 2016

[24] T T Nguyen J S Pan S C Chu J F Roddick and T K DaoldquoOptimization localization in wireless sensor network basedon multi-objective firefly algorithmrdquo Journal of NetworkIntelligence vol 1 no 4 pp 130ndash138 2016

[25] S Arora and S Singh ldquoNode localization in wireless sensornetworks using butterfly optimization algorithmrdquo ArabianJournal for Science and Engineering vol 42 no 8pp 3325ndash3335 2017

[26] S R Sujatha and M Siddappa ldquoNode localization method forwireless sensor networks based on hybrid optimization ofparticle swarm optimization and differential evolutionrdquo IOSR

Journal of Computer Networks and Communications 11

Journal of Computer Engineering vol 19 no 2 pp 07ndash122017

[27] R Kaur and S Arora ldquonature inspired range based wirelesssensor node localization algorithmsrdquo International Journal ofInteractive Multimedia and Artificial Intelligence vol 4pp 7ndash17 2017

[28] A +arwat E H Houssein M M Ahmed A E Hassanienand T Gabel ldquoMOGOA algorithm for constrained and un-constrained multi-objective optimization problemsrdquo AppliedIntelligence vol 48 no 8 pp 2268ndash2283 2017

[29] A A Ewees M Abd Elaziz and E H Houssein ldquoImprovedgrasshopper optimization algorithm using opposition-basedlearningrdquo Expert Systems with Applications vol 112pp 156ndash172 2018

[30] A G Hussien E H Houssein and A E Hassanien ldquoA binarywhale optimization algorithm with hyperbolic tangent fitnessfunction for feature selectionrdquo in Proceedings of Eighth In-ternational Conference on Intelligent Computing and In-formation Systems (ICICIS) pp 166ndash172 Cairo EgyptDecember 2017

[31] B Xue M Zhang and W N Browne ldquoParticle swarm op-timization for feature selection in classification a multi-objective approachrdquo IEEE Transactions on Cyberneticsvol 43 no 6 pp 1656ndash1671 2013

[32] S Arora and S Singh ldquoButterfly algorithmwith levy flights forglobal optimizationrdquo in Proceedings of IEEE InternationalConference in Signal Processing Computing and Control(ISPCC) pp 220ndash224 Waknaghat India September 2015

[33] S Arora and S Singh ldquoAn improved butterfly optimizationalgorithm with chaosrdquo Journal of Intelligent and Fuzzy Sys-tems vol 32 no 1 pp 1079ndash1088 2017

[34] R B Blair and A E Launer ldquoButterfly diversity and humanland use species assemblages along an urban grandientrdquoBiological Conservation vol 80 no 1 pp 113ndash125 1997

[35] X S Yang and X He ldquoFirefly algorithm recent advances andapplicationsrdquo 2013 httpsarxivorgabs13083898

[36] X S Yang ldquoFirefly algorithms for multimodal optimizationrdquoin Proceedings of International Symposium on Stochastic Al-gorithms pp 169ndash178 Sapporo Japan October 2009

[37] S Mirjalili S M Mirjalili and A Lewis ldquoGrey wolf opti-mizerrdquo Advances in Engineering Software vol 69 pp 46ndash612014

[38] L P Madin ldquoAspects of jet propulsion in Salpsrdquo CanadianJournal of Zoology vol 68 no 4 pp 765ndash777 1990

[39] S Mirjalili A H Gandomi S Z Mirjalili S Saremi H Farisand S M Mirjalili ldquoSalp swarm algorithm a bio-inspiredoptimizer for engineering design problemsrdquo Advances inEngineering Software vol 114 pp 163ndash191 2017

[40] P A Anderson and Q Bone ldquoCommunication betweenindividuals in Salp chains II Physiologyrdquo Proceedings of theRoyal Society of London Series B Biological Sciences vol 210no 1181 pp 559ndash574 1980

[41] H T Ibrahim W J Mazher O N Ucan and O BayatldquoFeature selection using salp swarm algorithm for real bio-medical datasetsrdquo IJCSNS vol 17 no 12 pp 13ndash20 2017

12 Journal of Computer Networks and Communications

Journal of Computer Engineering vol 19 no 2 pp 07ndash122017

[27] R Kaur and S Arora ldquonature inspired range based wirelesssensor node localization algorithmsrdquo International Journal ofInteractive Multimedia and Artificial Intelligence vol 4pp 7ndash17 2017

[28] A +arwat E H Houssein M M Ahmed A E Hassanienand T Gabel ldquoMOGOA algorithm for constrained and un-constrained multi-objective optimization problemsrdquo AppliedIntelligence vol 48 no 8 pp 2268ndash2283 2017

[29] A A Ewees M Abd Elaziz and E H Houssein ldquoImprovedgrasshopper optimization algorithm using opposition-basedlearningrdquo Expert Systems with Applications vol 112pp 156ndash172 2018

[30] A G Hussien E H Houssein and A E Hassanien ldquoA binarywhale optimization algorithm with hyperbolic tangent fitnessfunction for feature selectionrdquo in Proceedings of Eighth In-ternational Conference on Intelligent Computing and In-formation Systems (ICICIS) pp 166ndash172 Cairo EgyptDecember 2017

[31] B Xue M Zhang and W N Browne ldquoParticle swarm op-timization for feature selection in classification a multi-objective approachrdquo IEEE Transactions on Cyberneticsvol 43 no 6 pp 1656ndash1671 2013

[32] S Arora and S Singh ldquoButterfly algorithmwith levy flights forglobal optimizationrdquo in Proceedings of IEEE InternationalConference in Signal Processing Computing and Control(ISPCC) pp 220ndash224 Waknaghat India September 2015

[33] S Arora and S Singh ldquoAn improved butterfly optimizationalgorithm with chaosrdquo Journal of Intelligent and Fuzzy Sys-tems vol 32 no 1 pp 1079ndash1088 2017

[34] R B Blair and A E Launer ldquoButterfly diversity and humanland use species assemblages along an urban grandientrdquoBiological Conservation vol 80 no 1 pp 113ndash125 1997

[35] X S Yang and X He ldquoFirefly algorithm recent advances andapplicationsrdquo 2013 httpsarxivorgabs13083898

[36] X S Yang ldquoFirefly algorithms for multimodal optimizationrdquoin Proceedings of International Symposium on Stochastic Al-gorithms pp 169ndash178 Sapporo Japan October 2009

[37] S Mirjalili S M Mirjalili and A Lewis ldquoGrey wolf opti-mizerrdquo Advances in Engineering Software vol 69 pp 46ndash612014

[38] L P Madin ldquoAspects of jet propulsion in Salpsrdquo CanadianJournal of Zoology vol 68 no 4 pp 765ndash777 1990

[39] S Mirjalili A H Gandomi S Z Mirjalili S Saremi H Farisand S M Mirjalili ldquoSalp swarm algorithm a bio-inspiredoptimizer for engineering design problemsrdquo Advances inEngineering Software vol 114 pp 163ndash191 2017

[40] P A Anderson and Q Bone ldquoCommunication betweenindividuals in Salp chains II Physiologyrdquo Proceedings of theRoyal Society of London Series B Biological Sciences vol 210no 1181 pp 559ndash574 1980

[41] H T Ibrahim W J Mazher O N Ucan and O BayatldquoFeature selection using salp swarm algorithm for real bio-medical datasetsrdquo IJCSNS vol 17 no 12 pp 13ndash20 2017

12 Journal of Computer Networks and Communications

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom