ashippickupanddeliverymodelwithmul8ple commodies ...martrans.org › documents › 2012 › odysseus...
TRANSCRIPT
![Page 1: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/1.jpg)
A Ship Pickup and Delivery Model with Mul8ple Commodi8es, Variable Speeds, Cargo Inventory
Costs and Freight Rates
Harilaos N. Psara,is
Laboratory for Mari3me Transport Na3onal Technical University of Athens
Athens, Greece
![Page 2: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/2.jpg)
A synthesis of:
• Recent research in the area of pickup and delivery problems
• Recent research in the area of speed op3miza3on in mari3me transporta3on
• Recent research in the area of emissions from ships
Odysseus 2012 2
![Page 3: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/3.jpg)
The problem • Graph G(N,A) with N={0,1,2,..,n} a set of ports • Known inter-‐port distances sij • Known O/D matrix, amount of cargo from i to j [dij] • Each cargo a dis3nct commodity • Cargoes cannot be split (as many as n(n-‐1))
• Ship is ini3ally located at node 0 (home port) • Capacity Q ≥ max (i,j) dij • Has to pick up and deliver all cargoes, and return to port 0 • Port dwell 3mes propor3onal to cargo handled • Visit each port as many 3mes as necessary
Odysseus 2012 3
![Page 4: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/4.jpg)
O/D matrix
Odysseus 2012 4
![Page 5: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/5.jpg)
Cost func3on
• Fuel costs
• Time charter costs • Cargo inventory costs
Odysseus 2012 5
![Page 6: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/6.jpg)
Fuel costs
• On a leg from A to B of distance L
• If ship speed is v (n. miles/day)
• Fuel cost = PFUEL*(L/v)*FC(v)
• Where FC(v) is the ship’s daily fuel consump3on
Odysseus 2012 6
![Page 7: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/7.jpg)
Fuel costs
• FC = kV3 (cubic) • Reasonable approxima3on in many cases
• Problem: exponent may be >3 • Problem: FC=0 for v=0
Odysseus 2012 7
![Page 8: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/8.jpg)
More general FC
• FC = a+bVn (n≥3)
• Problem: FC depends on ship’s loading condition
Odysseus 2012 8
![Page 9: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/9.jpg)
Even more general FC • FC = (A+BVn)Δ2/3
Δ= ship’s displacement
• FC =f(V,w) (general)
• Depends on speed V and payload w
Odysseu
s 201
2 9
![Page 10: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/10.jpg)
Time charter costs
• Assume ship on 3me charter
• Time charter rate F ($/day)
• F exogenous, determined by market condi3ons
• Cost propor3onal to overall 3me of trip (which depends on speeds of ship on each leg of route)
Odysseus 2012 10
![Page 11: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/11.jpg)
Cargo inventory costs – Due to delay in delivery of cargo – Per unit volume and per unit 3me cargo delay cost is equal to α for cargo wai3ng to be picked up (cost accrues from 3me 0 un3l cargo is on the ship)
– Such cost is equal to β for cargo within the ship (cost accrues from 3me cargo is on the ship un3l cargo is delivered).
– Both α and β are constants ≥0. – Usually α=0 for JIT systems – These costs can be important if inventory costs are important (mainly for long-‐haul problems and/or high valued cargoes)
Odysseus 2012 11
![Page 12: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/12.jpg)
What are α and β?
• Lower bound in both α and β is PR/365
• Where P is CIF value of cargo • R is cargo owner’s cost of capital • (α, β high for expensive cargoes)
Odysseus 2012 12
![Page 13: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/13.jpg)
Important observa3on • Ship speed impacts all three categories of costs
• Fuel costs in a posi3ve way • Time charter costs in a nega3ve way • Cargo inventory costs in a nega3ve way • What should the set of op3mal ship speeds be? • What should be the rou3ng? • What should be the sequence of pickups and deliveries?
Odysseus 2012 13
![Page 14: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/14.jpg)
Role of speed
• Has always been important
• Increasingly important in recent years
• Economic considera3ons
• Environmental considera3ons
Odysseus 2012 14
![Page 15: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/15.jpg)
*Psara,is, H.N. and C.A. Kontovas (2009), “CO2 Emissions Sta3s3cs for the World Commercial Fleet”, WMU Journal of Mari3me Affairs, 8:1, pp. 1-‐25.
Odysseus 2012 15
![Page 16: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/16.jpg)
Speed reduc3on
• An obvious way to reduce emissions
• Killing 3 birds with one stone?
• Pay less for fuel • Reduce CO2 (and other) emissions • Help sustain a vola3le market
Odysseus 2012 16
![Page 17: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/17.jpg)
Dual targesng
• OPERATIONAL • Operate exis3ng ships at reduced speed (derate engines)
• Slow steaming kits
• STRATEGIC (DESIGN)
• Design new ships that cannot go very fast (have smaller engines)
Odysseus 2012 17
![Page 18: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/18.jpg)
How much slower?
• From 20-‐25 knots, go down to 14-‐18
• New Maersk 18,000 TEU ships: 19 knots
• Project ULYSSES: Go 5-‐6 knots!
Odysseus 2012 18
![Page 19: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/19.jpg)
Some basics
• Ships do NOT trade at predetermined speeds. • Those who pay for the fuel, that is, the ship owner if the
ship is in the spot market on voyage charter, or the charterer if the ship is on time or bareboat charter, will choose an optimal speed as a function of
– (a) bunker price, and – (b) the state of the market and specifically the spot rate
Odysseus 2012 19
![Page 20: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/20.jpg)
Odysseus 2012 20
![Page 21: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/21.jpg)
Basics ii
• Even though the owner’s and time charterer’s speed optimization problems may seem at first glance different, for a given ship the optimal speed (and hence fuel consumption) is in both cases the same.
• In that sense, it makes no difference who is paying for the fuel, the owner, the time charterer, or the bareboat charterer.
Odysseus 2012 21
![Page 22: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/22.jpg)
Owner in spot market
• OBJECTIVE: Maximize average per day profits • s: spot rate ($/tonne) • C: payload (tonnes) • p: fuel price • F(v): fuel consumption at speed v • D: route r-trip distance • E: OPEX ($/day)
Odysseus 2012 22
![Page 23: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/23.jpg)
Time charterer • OBJECTIVE: Minimize average per day costs • R: demand requirements (tonnes/day) • T: time charter rate ($/day)
Odysseus 2012 23
![Page 24: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/24.jpg)
Role of ra3o ρ= p/s
• Both problems reduce to:
Odysseus 2012 24
![Page 25: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/25.jpg)
Ra3o ρ=p/s
Odysseus 2012 25
![Page 26: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/26.jpg)
Taxonomy of speed models
• Psara,is & Kontovas (2012)
• Non-‐emissions related
• Emissions-‐related
Odysseus 2012 26
![Page 27: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/27.jpg)
Classifica3on according to
• Op3miza3on criterion: cost, profit, or other • Shipping market/context • Who is the decision maker • Fuel price an input? • Freight rate an input? • Fuel consump3on func3on? Cubic/general • Op3mal speeds in various legs • Logis3cal context
Odysseus 2012 27
![Page 28: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/28.jpg)
Classifica3on ii
• Size of fleet? Single ship, mul3ple ships • Adding more ships an op3on? • Inventory costs included? • Emissions considered? • Modal split considered? • Ports included in formula3on?
Odysseus 2012 28
![Page 29: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/29.jpg)
Sample output
Odysseus 2012 29
22
TABLE 3a: Taxonomy part I
Taxonomy parameter \ paper Alderton (1981) Bausch et al (1998) Benford (1981) Brown et al
(1987) Cariou (2011) Cariou and Cheaitou
(2012)
Corbett et al (2010)
Devanney (2007)
Devanney (2010)
Eefsen and Cerup-
Simonsen (2010)
Optimization criterion Profit Cost Cost Cost Cost Cost Profit Profit Cost or profit Cost
Shipping market General Tanker/ barge Coal Tanker Container Container Container
Tanker Tanker (VLCC) Container
Decision maker Owner Owner Owner Owner Owner Owner Owner Owner Either Owner Fuel price an explicit input Yes Yes No No Yes Yes Yes Yes Yes Yes Freight rate an input Input No No No No No Input Computed Computed No Fuel consumption function Cubic Unspecified Cubic Unspecified Cubic Cubic Cubic Cubic General Cubic Optimal speeds in various legs Yes No No Only ballast No No No Yes Yes No
Optimal speeds as function of payload Yes No No No No No No No No No
Logistical context Fixed route Routing and scheduling
Fleet deployment
Routing and scheduling Fixed route Fixed route Fixed route World oil
network Fixed route Fixed route
Size of fleet Multiple ships Multiple ships Multiple ships Multiple ships Multiple ships Multiple ships Multiple ships Multiple ships One ship Multiple ships Add more ships an option Yes No No No Yes Yes Yes Yes Yes Yes Inventory costs included Yes No No No No Yes No Yes Yes Yes Emissions considered No No No No Yes Yes Yes No No Yes Modal split considered No No No No No No No No No No Ports included Yes Yes No No No Yes No Yes No Yes
TABLE 3b: Taxonomy part II
Taxonomy parameter \ paper Faber et al (2010)
Fagerholt (2001)
Fagerholt et al (2010)
Gkonis Psaraftis
(2011abcd)
Kontovas Psaraftis (2011)
Lindstad et al (2011)
Norstad et al (2011)
Notteboom Vernimmen
(2010)
Papadakis Perakis (1989) Perakis (1985)
Optimization criterion No/A Cost Cost Profit Cost Pareto analysis Cost Cost Cost Cost
Shipping market Various General Liner Tanker, LNG, LPG Container All major ship
types Tramp Container Tramp Tramp
Decision maker No/A Owner Owner Owner Charterer Owner Owner Owner Owner Owner Fuel price an explicit input No No No Yes Yes Yes No Yes Yes No Freight rate an input No No No Input Input No No No No No Fuel consumption function Cubic Cubic Cubic General Cubic Cubic Cubic Unspecified General Cubic Optimal speeds in various legs No Yes Yes Yes Yes No Yes No Yes No
Optimal speeds as function of payload No No No No Yes Yes No No No No
Logistical context Fixed route Pickup and delivery Fixed route Fixed route Fixed route Fixed route Pickup and
delivery Fixed route Fleet deployment
Fleet deployment
Size of fleet Multiple ships One ship One ship Multiple ships Multiple ships Multiple ships Multiple ships Multiple ships Multiple ships Multiple ships Add more ships an option Yes No No Yes Yes Yes No Yes No Yes Inventory costs included No No No Yes Yes Yes No No Yes No Emissions considered Yes No Yes Yes Yes Yes No No No No Modal split considered No No No No No No No No No No Ports included No No No Yes Yes Yes No Yes Yes No
TABLE 3c: Taxonomy part III
![Page 30: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/30.jpg)
VLCC speed model
• Gkonis & Psara,is (2012)
Odysseus 2012 30
DATA
EMISSIONS CALCULATIONS(CO2, NOx, SO2, PM)
EMISSIONS GRAPHS
TRADE ROUTE
FUELS
COSTS - FREIGHT
MAIN ENGINE
1-vessel CALCULATIONS
INDEX - TABLE OF CONVERSIONS
SPEED OPTIMISERENUMERATION OF RESULTS -
OPTIMAL SPEEDS
![Page 31: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/31.jpg)
VLCC results • Route: Gulf-‐Japan • Op3mize both laden and ballast
speeds
Odysseus 2012 31
![Page 32: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/32.jpg)
VLCC cont’d • Include cargo inventory costs
Odysseus 2012 32
![Page 33: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/33.jpg)
Effect of fuel price on emissions
•
Odysseus 2012 33
![Page 34: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/34.jpg)
Speed decision can be decomposed from rou3ng decision
• Assuming the ship is at port A and is set to sail to port B, the total cost on leg (A, B) is equal to
• COST(A,B) = [PFUELf(v, w) + αu + βw + F](sAB /v),
Where: • v: ship speed during leg • w: ship payload during leg • u: total weight of cargo not yet picked up during leg
Odysseus 2012 34
![Page 35: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/35.jpg)
Decompose speed cont’d • Factor out sAB
• INCR(A,B) = min v∈S {[PFUELf(v, w) + αu + βw + F]/v}, with with S={v: vLB(w) ≤ v ≤ vUB(w)}
(per mile total cost)
• Observa3on: Speed decision is independent of A or B
Odysseus 2012 35
![Page 36: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/36.jpg)
2nd observa3on
• Input parameters PFUEL, F, α, and β are key determinants of the speed decision
• Higher values of PFUEL would reduce op3mal speed
• Higher values of F, α, or β would increase op3mal speed
Odysseus 2012 36
![Page 37: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/37.jpg)
3rd observa3on
• Input parameters PFUEL, F, α, and β can also influence the ROUTING decision!
Odysseus 2012 37
![Page 38: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/38.jpg)
Example: ship of Q=11 (000 tons)
Odysseus 2012 38
![Page 39: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/39.jpg)
Minimum fuel cost (F=α=β=0)
• v between 8 and 14 knots
• Cubic FC func3on • FC dependence on w • Fuel price $600/ton
• Sail at minimum speed • Op3mal route: 0-‐1-‐2-‐3 • even though total distance sailed (660 nau3cal miles) is more than that of route 0-‐2-‐1-‐3 (480 nau3cal miles).
• Reason: heavier cargo is delivered first
Odysseus 2012 39
![Page 40: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/40.jpg)
If F > $450/day
• Op3mal route: 0-‐2-‐1-‐3 • Different speeds in each leg
• Speeds depend on F (higher if F increases)
Odysseus 2012 40
![Page 41: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/41.jpg)
Embed within pickup and delivery algorithm
• Extension of:
Odysseus 2012 41
![Page 42: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/42.jpg)
Special case 1: dial-‐a-‐ride problem
• Psara,is (1980, 1983abc), Desrosiers et al (1986), Cordeau (2006), Cordeau et al (2008)
• dij=1 for j=i+n only, 0 otherwise
Odysseus 2012 42
![Page 43: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/43.jpg)
Special case 2: vehicle rou3ng problem with pickup and delivery
• Kalantari et al (1985), Desrosiers et al (1986), Ruland and Rodin (1997), Cordeau et al (2008), Berbeglia et al (2007).
• dij ≠0 for j=i+n only, 0 otherwise •
Odysseus 2012 43
![Page 44: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/44.jpg)
Special case 3: all nodes feed into a depot and depot sends to all nodes
• Gribkovskaia et al (2006) • Depot at one node Z
Odysseus 2012 44
![Page 45: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/45.jpg)
Special case 4: unidirec3onal
• Hernandez-‐Perez and Salazar-‐Gonzalez (2009) • If cargo from A to B, no cargo from B to A
Odysseus 2012 45
![Page 46: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/46.jpg)
DP approach
Odysseus 2012 46
![Page 47: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/47.jpg)
Recursion
Odysseus 2012 47
![Page 48: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/48.jpg)
Status
• Developed combined algorithm • Tested it on sample problems
• Developed NON-‐ROUTING speed models for tankers, bulk carriers, ro/ro carriers, LNG/LPG, container vessels
• Plans to integrate these models to fleet case
Odysseus 2012 48
![Page 49: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/49.jpg)
Odysseus 2012 49
![Page 50: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/50.jpg)
Which model?
• Long haul
Odysseus 2012 50
![Page 51: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/51.jpg)
Short haul (if price of emissions is high enough)
Odysseus 2012 51
![Page 52: AShipPickupandDeliveryModelwithMul8ple Commodies ...martrans.org › documents › 2012 › Odysseus Psaraftis.pdf · Sample!output Odysseus2012 ! 29 22 TABLE 3a: Taxonomy part I](https://reader036.vdocuments.net/reader036/viewer/2022081612/5f12f08d025ecc2981520b14/html5/thumbnails/52.jpg)
Is this green enough?
• Globally, ruminant livestock produce about 80 million metric tons of CH4 annually, accoun3ng for about 28% of global CH4 emissions from human-‐related ac3vi3es (source: US EPA)
Odysseus 2012 52