ALGORITHM FOR TRANSPORTATION ROUTING AND VEHICLE LOADING

A model is formulated and a computationally efficient heuristic procedure for its solution is presented and illustrated. The methodology is applicable to large size problems with multiple stops, commodities, modes of transportation and time periods. Individual vehicles can be routed through a time dimensioned network with the integer nature of the routes preserved. An individual vehicle can carry more than one commodity and trans-shipment is permitted. Constraints on vehicle, node and arc capacities are represented. The objective function can include both operating costs and penalties on early or late arrivals at destinations. Solutions are obtained using a modification of Dantzig-Wolfe decomposition termed "reflection" programming. Large problems require only a few minutes for solution.