INTEGER PROGRAM OPTIMIZATION OF COACH ASSIGNMENT WITHIN AN OVERLAPPED NETWORK

Abstract

In these days of increasing traffic congestion, increasing energy prices, and decreasing transportation funding it is imperative that efficient, alternate transportation be maintained. It is therefore the goal of this thesis to propose an Integer Program model for optimizing train consists (the number of cars assigned to a particular passenger train) to lower the operational costs while still meeting demand. Further benefits are the increased utilization of the existing car fleet of the service optimized and the reduction of the overall car fleet required. All of these goals are met by the model contained here-in, and validated through an optimization of Amtrak's Northeast Operations. The model shows distinct improvements in lowering operational costs, reducing the overall fleet required, and increasing car utilization for all cases optimized. These include cases to determine sensitivity analysis, where a minimum train length is imposed and where a maximum terminal capacity is imposed.