Optimization of a Transportation System Planning Problem.

摘要

This study investigated the problem of synthesizing a minimum cost transportation system plan to service forecast shipment requirements among a set of points in such a manner as to satisfy aggregate ship-time performance levels. Both commercial and dedicated modes may be used in the transportation plan; but the later must be designed in detail including specification of vehicles to be used, the route each is to service, and arrival/departure time schedules. A large-scale, mixed integer, linear programming model of the planning problem is developed and simplified for solution by applying Benders' decomposition to yield two more simple, interacting subproblems. One of these, a linear program which assures ship-time performance, is amenable to large-scale programming techniques for which specialized algorithms are stated. Several formulations of the other subproblem, which defines the dedicated mode network, are described. The first designs the network using a set of feasible, vehicle tours. An implicit enumeration algorithm applicable to this model is described. Three additional formulations, each of which constructs vehicle tours directly, were developed and solution approaches for each are described. The tactical, vehicle scheduling problem is treated subsequently. Collectively, study results offer capability to solve transportation system planning problems of realistic size. (Author)