1. Cooperation in Urban Transportation Tom Van Woensel

2. Menu for today Urban environments Horizontal collaboration between LSPs Analysis and discussion Planning and scheduling of their operations: Vehicle Routing Problems Gain sharing mechanisms Collaboration between public and freight transport Caveat emptor: ongoing work ! / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 2

3. Urban Environments and Transportation Transportation demand is growing and expected to continue growing Last-mile logistics is more and more situated in growing urban areas: Increasing urban population growth (Over)congested Difference with emerging cities Both internal costs and the external costs are key logistics drivers Decarbonisation to reduce pollution / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 3

4. Urban Environments and Transportation Quality (e.g. carbon footprint) and quantity (e.g. transport movements) of the distribution activities Better orchestration of the different physical flows by considering all relevant stakeholders −Retailers −LSPs −Government −Public Coordination and consolidation challenges are not new, but sound and sustainable solutions are not easily realized. / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 4

5. (The lack of) Cooperation and coordination in urban transportation / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 5

6. Routing and Scheduling in urban areas / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 6

7. Two-index formulation for VRP / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 7

8. VRP and Time windows / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 8 =1 if vehicle k drives directly from i to j The time vehicle k starts Servicing customer j cost trav. from i to j demand at customer i capacity of vehicle k depot at start depot at end time to travel directly from i to j plus service time in i large number ASSIGNMENT CAPACITY FLOW CONSER- VATION TIME WINDOWS

9. dashboard LSP 1 Parameters 1.> 3plts 2.north NL 3.HACCP 4.Non smelling 5. …. Parameters 1.unrestricted 2.LZV only 3.TLN member 4.Digital VVA 5.…. Parameters 1.FTL only 2.Digital VVA 3.chilled 4.…. dashboard LSP 2 dashboard LSP 3 Three Dutch carriers / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 9

10. Three carriers consolidated DeltaAbsolute Δ Cost-factor-4.8%-24.993 Δ Kms driven-4.7%-32.463 Δ Empty kilometers-15.2%-38.547 Δ Driving hours-4.9%-497 Δ Vehicles-12.9%-55 # Orders6.762 # Routes3.647 # Pallets83.887 # Vehicles as-is425 Gain sharing mechanisms needed / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 10

11. Collaboration among Logistics Service Providers / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 11 DistanceNumber of TrucksUtilization (avg) LSP 1764.791046% LSP 2750.28949% Total1515.071947% Cooperation828.941075% Difference-45%-47%+38%

12. Some variants to the problem DistanceVehicles used LSP1LSP2TotalLSP1LSP2Total Equal (50/50)764.79750.281515.0791019 Unequal (20/80)590.57790.871381.4461016 Regional (C/R) No overlap 650.94217.30868.248311 Regional (C/R) Minor overlap 642.98297.09940.078412 / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 12 Cooperation: Distance = 828.94 with 10 vehicles C/R = Country/Regional player

13. What about the cost savings? / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 13 Total Km savingsKm savings per drop LSP1LSP2TotalLSP1LSP2Total Equal (50/50)255.49256.70512.185.115.135.12 Unequal (20/80)316.0295.20411.2215.801.194.11 Regional (C/R) No overlap 0.0017.76 0.000.660.18 Regional (C/R) Minor overlap 48.2823.2071.480.690.770.71 Cost savings are not always equal over all partners Limited savings if no synergy in the networks

14. Shapley value for player i: Allocates to each participant its average marginal contribution Complete random order of entering of participants For two players: Equal allocation of the total value of cooperation Regardless of the specific characteristics (e.g. size, orders, etc.) of the two players How to share the benefits? / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 14

15. Side Payments are needed / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 15 Total km savingsSide payments LSP1LSP2TotalLSP1LSP2 Equal (50/50)255.49256.70512.18-0.610.61 Unequal (20/80)316.0295.20411.22110.41-110.41 Regional (G/R) No overlap 0.0017.76 -8.888.88 Regional (G/R) Minor overlap 48.2823.2071.48-12.5412.54 A transparent and clear cost accounting system is needed

16. What about shippers? Tariffs adapted following LSP collaborations Proposal: follow Shapley allocation Two players: Shipper and LSP 50/50 allocation rule / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 16 No collaborationCollaboration – no sharing with Shipper Sharing with Shipper Tariff100 95 (-5 from coll.) Costs9790 Margin3105 How certain are we about this?

17. What about time and stochasticity? Volumes offered to the coalition Changes over time, contracts, tenders, etc. How to cope with this if one partner falls behind? Consequences are significant / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 17 No collaborationSharing with Shipper Volumes in collaboration  Tariff10095 (-5 from coll.)95 Costs9790[92,97] Margin35[3,-2]

18. What do we need? Sound concepts for gain sharing: That survive over time and uncertainty That are fast measurable/quantifiable throughout the collaborations That consider the dynamics and stochastics of the LSP and shipper’s world A gain sharing control tower? Neutral instance Open bookkeeping with regards to costs / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 18

19. A different cooperation Combine passenger transport and freight transport Different levels: Share dedicated resources, e.g. priority lanes Share equipment: taxi, bus, train, boat Some examples: Passenger airplanes also carry freight Norwegian Hurtigruten: mail, cargo and passengers Dabbawalla / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 19

20. Dabbawalla 0% fuel, 0% investment, 0% modern tech, 0% Disputes, 99.99% performance, 100 % Customers Satisfaction 400,000 transactions every day (including return), i.e. 400,000*25 days*12 months= 120,000,000 transactions per year / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 20

21. Grant Road (12) Churchgate (1-10) Lower Parel (14) 2 3 7 4 5 6 Zones for destination Distribution By Carriers at lunchtime To offices Point of Aggregation And Sorting C DB A E Collection from home 1 The Flow Logic: using trains! / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 21

22. Using taxi’s for parcel deliveries / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 22

23. Trail of one taxi in San Francisco (1 day) / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 23

24. Trail of 30 taxis in San Francisco (1 day) / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 24

25. Taxis for parcel delivery Online optimization Stochastic customers Stochastic demand Pickup and delivery problems Using taxis for parcels: Utilized (with customer): time value? Available (no customer): use the ‘dead’ time Joint work started with Connexxion for The Netherlands PhD research of Baoxiang Li / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 25

26. Overview and summary Collaboration among LSP Gain sharing principles How to translate this into practice? The role of shippers in the collaborations Collaboration between public and freight transport Interesting research questions arise Still an open research field / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 26

27. Thank you. Any questions/suggestions/com ments? / School of Industrial Engineering - prof. dr. Tom Van Woensel PAGE 27