![Page 1: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/1.jpg)
Femtocell Base Station Deployment in Commercial Buildings:
A Global Optimization Approach
Adviser: Frank , Yeong - Sung Lin
Present by Li Wen Fang
![Page 2: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/2.jpg)
2
Agenda
• Introduction• Network Model and Problem Formulation• A Global Optimization Solution Approach• Numerical Results• Conclusion
![Page 3: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/3.jpg)
3
Agenda
• Introduction• Network Model and Problem Formulation• A Global Optimization Solution Approach• Numerical Results• Conclusion
![Page 4: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/4.jpg)
4
Introduction• Femtocells are low-cost, small-sized base stations (BS) that
offer high quality voice/data services to indoor mobile handset users.
• “Win-win” situation - not only do users enjoy better coverage due to the close vicinity to BSs, the network operators also benefit from a reduced demand for constructing macrocell towers.
• The special features of commercial building environments bring many new challenges to femtocell deployments.
• Femtocell placement optimization in commercial buildings is still under-explored.
![Page 5: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/5.jpg)
5
Introduction• Femtocell placement optimization in commercial buildings is
important due to the high impact of the locations of femtocell base stations (FBS) on the energy expenditure of each mobile handset (HS).
• Transmission power of an HS depends heavily on the physical distance between the HS and its targeted base station.
• Unlike simple residential buildings, commercial buildings range from small offices to large-sized shopping malls with different building layouts and various wall and floor, which yield different signal path losses.
![Page 6: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/6.jpg)
6
Introduction• We focus on the joint optimization of FBS placement to
minimize the uplink transmission power of each HS.
• Main contributions:
mathematical model, unique floor attenuation factor, FBS installation restrictions
transform the original MINLP problem into an equivalent MICP problem
a global optimization approach based on its convex programming relaxation (CPR) and the branch-and-bound (BB) framework
![Page 7: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/7.jpg)
7
Agenda
• Introduction• Network Model and Problem Formulation• A Global Optimization Solution Approach• Numerical Results• Conclusion
![Page 8: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/8.jpg)
8
Network Model and Problem Formulation
A. Network Model
B. Path Loss Modeling for Commercial Buildings
C. QoS Requirement Constraints
D. FBS Association Modeling
E. Problem Formulation
![Page 9: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/9.jpg)
9
A. Network Model• We consider a femtocell network in a commercial building
consisting of M FBSs and N service areas (SA). As illustrated in Fig. 1.
• Distance relationship between an FBS and an SA
– Horizontal distance between FBS m and SA i, as shown in Fig. 2.
– Vertical distance, as shown in Fig. 3.
mounted on the ceiling
be integer-valued
![Page 10: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/10.jpg)
10
A. Network Model• The distance between FBS m and SA i
![Page 11: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/11.jpg)
11
B. Path Loss Modeling for Commercial Buildings
• For the path loss within the same floor, we adopt the following equation to model path loss (in dBm)
• Next, we incorporate the path loss that result from floor separations.
![Page 12: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/12.jpg)
12
• Lemma 1.
Under the wireless signal path loss model in (3) and (4) and upon converting PRm, PTi, and Ld0 to a suitable linear scale, the following relationship holds:
B. Path Loss Modeling for Commercial Buildings
![Page 13: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/13.jpg)
13
• Satisfies the HS’s QoS (Quality of Service) requirement. Based on the path loss model in Lemma 1
• By rearranging terms and letting
C. QoS Requirement Constraints
![Page 14: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/14.jpg)
14
• Rewrite the QoS constraints in (6)
C. QoS Requirement Constraints
![Page 15: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/15.jpg)
15
• For each SA, the channel to the nearest FBS may or may not be the best because the nearest FBS could be separated by a floor.
From the path loss model in (5), it is not difficult to see that the nearest FBS might actually have a worse path loss because of the FAF effect.
• Then, the FBS association can be modeled as
D. FBS Association Modeling
![Page 16: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/16.jpg)
16
• Due to the above FBS association modeling technique, we need to modify the QoS constraints in (7) as follows:
D. FBS Association Modeling
![Page 17: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/17.jpg)
17
• To prolong the HS battery lives and ensure fairness among all SAs, we can minimize the maximum uplink transmission power originating from the SAs.
• FBS placement and power control problem (FPPC)
E. Problem Formulation
![Page 18: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/18.jpg)
18
E. Problem Formulation
![Page 19: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/19.jpg)
19
Agenda
• Introduction• Network Model and Problem Formulation• A Global Optimization Solution Approach• Numerical Results• Conclusion
![Page 20: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/20.jpg)
20
A Global Optimization Solution Approach
A. Problem Reformulation
B. A Solution Procedure Based on a Branch-and-Bound Framework and Convex Programming Relaxations (BB/CPR)
C. Convergence Speedup: A Symmetry-Defeating Approach
![Page 21: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/21.jpg)
21
• Our ultimate goal is to convert the highly unstructured constraint (13) and the nonconvex constraint (15) into a form that is solvable by standard optimization tools.
• Reformulating the Distance Constraint in (15) 1
Accordingly, (13) becomes
A. Problem Reformulation
![Page 22: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/22.jpg)
22
• Lemma2.
Constraint (16) can be equivalently replaced by
A. Problem Reformulation
![Page 23: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/23.jpg)
23
• Then, (18) can be rewritten as the following group of constraints:
A. Problem Reformulation
![Page 24: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/24.jpg)
24
• Lemma 3.
The distance constraint (15) can be convexified as:
A. Problem Reformulation
![Page 25: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/25.jpg)
25
• Reformulating the Minimum Received Power Constraint in (13)
• We linearize (17) with respect to the binary variables πim. To this end, we have the following lemma:
• Lemma 4.
Constraint (13) is equivalent to the following alternative representation:
A. Problem Reformulation
![Page 26: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/26.jpg)
26
• In Lemma 4, a valid value for the upper bound Uim can be chosen as
• Recalling that xmax, ymax, and F are the maximum length, width, and number of floors of the building, we can compute in (22) as:
A. Problem Reformulation
![Page 27: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/27.jpg)
27
• To further simplify (21), we introduce two new variables and rewrite (21) as the following three constraints:
• We can convexify the constraint (23-2) as: (following the same argument as in Lemma 2)
A. Problem Reformulation
![Page 28: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/28.jpg)
28
• Since Zm is integer-valued on { 1, . . . , F } , we can rewrite it using the following equivalent binary representation:
Where λmk is binary and satisfies
A. Problem Reformulation
![Page 29: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/29.jpg)
29
• As a result, it is easy to verify the (23)-3 is equivalent to the following linear constraint in λmk -variables:
• Substituting (27) into the first constraint in (23), the latter becomes:
A. Problem Reformulation
![Page 30: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/30.jpg)
30
• Recall that A(Zm, Wi) is equal to A0, if Zm = Wi ; or equals A1, if Zm Wi.
Thus, (28) can be further written as
• So far, we have converted the highly unstructured expression in (13) to an expression in (29) that is linear in the binary variables πim but has bilinear terms νim λmk .
A. Problem Reformulation
![Page 31: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/31.jpg)
31
• First, we let
• Then, (29) can be linearized as
A. Problem Reformulation
![Page 32: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/32.jpg)
32
• Theorem 1.
Given (26) with λmk { 0, 1 }, m, k, the bilinear equation ∈ ∀(30) holds if and only if
A. Problem Reformulation
![Page 33: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/33.jpg)
33
• Putting together all the previous derivations, we obtain the following equivalent reformulation for FPPC:
A. Problem Reformulation
![Page 34: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/34.jpg)
34
• In R-FPPC, all constraints are either linear or convex, and so R-FPPC is a mixed-integer convex program (MICP), which is readily solvable using a branch-and-bound (BB) framework, where lower bounds are computed using the continuous convex programming relaxation (CPR) of R-FPPC.
A. Problem Reformulation
![Page 35: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/35.jpg)
35
• Using the CPR, we can efficiently compute a global lower bound, LB for R-FPPC.
• The BB process proceeds by tightening LB and UB through a series of partitions over the problem domain, and terminates when UB ≤ (1 + ϵ)LB is satisfied, where ϵ > 0 is some desired approximation error.
• The BB/CPR process is guaranteed to converge to a global optimal solution.
B. A Solution Procedure Based on BB/CPR
![Page 36: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/36.jpg)
36
(a) The original problem.
(b) The first partition.
(c) Further partition.
Fig. 4. Illustration of the branch-and-bound solution framework.
Bou
nds
for
obje
ctiv
e va
lue
Bou
nds
for
obje
ctiv
e va
lue
Bou
nds
for
obje
ctiv
e va
lue
![Page 37: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/37.jpg)
37
![Page 38: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/38.jpg)
38
C. Convergence Speedup: A Symmetry-Defeating Approach
• There exists a symmetry with respect to the FBSs in the model.
• The solution (x1 , y1 , z1) = (50.2, 26.1, 2), (x2 , y2 , z2) = (75.2, 39.1, 2), and (x3, y3, z3) = (18.4, 56.3, 3) is essentially not different from (x3, y3, z3) = (50.2, 26.1, 2), (x2 , y2 , z2) = (75.2, 39.1, 2), and (x1 , y1 , z1) = (18.4, 56.3, 3).
• This phenomenon is due to the fact that the FBSs are indistinguishable objects by nature.
![Page 39: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/39.jpg)
39
C. Convergence Speedup: A Symmetry-Defeating Approach
• To defeat this symmetry, we introduce the following lexicographic order constraint:
![Page 40: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/40.jpg)
40
C. Convergence Speedup: A Symmetry-Defeating Approach
• Proposition 1.
Then, the constraint in (35) holds if and only if
![Page 41: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/41.jpg)
41
Agenda
• Introduction• Network Model and Problem Formulation• A global Optimization Solution Approach• Numerical Results• Conclusion
![Page 42: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/42.jpg)
42
Numerical Results• We first use a 20-SA network in a 3-story building as an
example. The building’s length, width, and per-floor height are 100, 60, and 3 meters, respectively.
• The length and width of all SAs are randomly generated and uniformly distributed in the intervals [0, 20] and [0, 16] (in meter).
• We plan to place three FBSs to serve this network. • The maximum transmission power limit and minimum
received power threshold for each HS are 1W and −100 dBm, respectively. The path loss exponent is 4.
![Page 43: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/43.jpg)
43
Fig. 5. The optimal FBS placement and FBS-SA association relationship for a 20-SA network.
• Due to FAF, not all SAs are associated with their nearest FBSs.
![Page 44: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/44.jpg)
44
TABLE 1The coordinates and sizes of SAS in the network in Fig. 5.
![Page 45: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/45.jpg)
45
TABLE 2The optimal FBS locations for the network in Fig. 5.
![Page 46: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/46.jpg)
46
TABLE 3The effect of symmetry-defeating on the speed of convergence.
![Page 47: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/47.jpg)
47
Fig. 6. The convergence process of BB/CPR for the network example in Fig. 5.
![Page 48: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/48.jpg)
48
Fig. 7. The scaling of convergence time with respect to the number of SAs. The number of FBSs is three.
![Page 49: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/49.jpg)
49
Fig. 8. The scaling of convergence time with respect to the number of FBSs. The number of SAs is 20.
![Page 50: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/50.jpg)
50
Fig. 9. The scaling of minimized maximum transmitted power (objective value) with respect to the number of SAs. The number of FBSs is three.
![Page 51: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/51.jpg)
51
Fig. 10. The scaling of minimized maximum transmitted power (objective value) with respect to the number of FBSs. The number of SAs is 10.
![Page 52: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/52.jpg)
52
Fig. 11. The scaling of average minimum required number of FBSs with respect to the shadowing effect deviation.
![Page 53: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/53.jpg)
53
Agenda
• Introduction• Network Model and Problem Formulation• A Global Optimization Solution Approach• Numerical Results• Conclusion
![Page 54: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/54.jpg)
54
Conclusion • The results in this paper offer both important analytical tools
and practical insights that enhance our understanding of femtocells deployments in commercial building environments.
• Possible future directions include the study of joint spectral management and placement optimization problems, and the study of combining FBS placement with multiple-antenna systems and downlink power control.
![Page 55: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/55.jpg)
55
Thanks for your listening !
![Page 56: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/56.jpg)
56
Fig. 1. An illustration of a femtocell network with multiple FBSs and SAs in a multi-floor commercial building. The FBSs are mounted on the ceiling of each floor.
![Page 57: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/57.jpg)
57
Fig. 2. The projected horizontal distance between FBS m and SA i.
• The furthest point in SA i away from FBS m is point C.
•
![Page 58: Femtocell Base Station Deployment in Commercial Buildings : A Global Optimization Approach Adviser: Frank, Yeong - Sung Lin Present by Li Wen Fang](https://reader038.vdocuments.net/reader038/viewer/2022110403/56649e6b5503460f94b68a4a/html5/thumbnails/58.jpg)
58
Fig. 3. The vertical distances between femtocell base stations and handsets.
• The height of each floor is h.
• Average height of an HS is η.