Transportation Logistics CEE 498 B599 I Professor Goodchild

Transportation Logistics CEE 498 B/599 I Professor Goodchild 4/18/07

Routing and Scheduling • Which customers to be visited by each vehicle? • Sequence in which they will be visited • When will each customer be visited? • How much will this cost? • Minimize cost and travel time, missed deliveries, comply with constraints (truck size, driver hours)

TSP and VRP • Traveling Salesman Problem (TSP): Given N points, find a tour that visits them all, returning to the point of departure, with minimum distance (here we are not dealing with many vehicles) • Vehicle Routing Problem (VRP): Find an allocation of points to vehicles, and a set of tours that return to a depot, that minimize either distance, number of vehicles, or a combination of the two (this is the same problem we addressed with the SMM and GA method, but number of vehicles not given)

Text Example: Peapod • • Online grocer Each day has different realization of deliveries Webvan Example: – distance on grid proportional to actual distance traveled – Cost proportional to distance traveled – Capacity 200 units – Minimize distance traveled

Location/Allocation • Segment the area (assign customers to a route) • Decide on route

Sequence Customers on Routes • Determine an initial route (not necessarily optimal) 1) Farthest insert (add farthest from DC first) 2) Nearest insert (add closest to DC first) 3) Nearest neighbor (add closest to customer just added first) 4) Sweep

Results are not always intuitive Farthest insert (add 56 farthest from DC first) Nearest insert (add closest to DC first) 56 Nearest neighbor (add closest to customer just added first) 66 Sweep 56

Route Improvement Procedures • Can we do better than original tours? • There are many ways you could do this: – 2 OPT: break trip 2 spots, reconnect – 3 OPT: break trip 3 spots, reconnect

Transportation Networks • • Direct Shipping Milk-run Distribution center (hub) Tailored network (serving different customers with different modes, methods or frequencies)

Trade-offs • • Management complexity Inbound/outbound transportation costs Cost of facility Inventory Handling Ability to respond to the customer Mode

In-class exercise • What attributes of a logistics system make milk-runs appropriate? (e. g. high density customers, large loads) • What attributes make a cross-docking facility appropriate? • When does it make sense to aggregate inventory?