NETWORK REDUCTION FOR THE ACYCLIC CONSTRAINED SHORTEST-PATH PROBLEM

This paper presents a procedure for reducing networks used with the constrained shortest path problem. The network considered herein is acyclic with each arc having two attributes; time and length, with a time limit on the shortest length path. The procedure traverses the network twice and depends on establishing time labels for the nodes similar to those established in shortest path algorithms of the label setting type. These labels are used to identify the arcs and nodes that render a path infeasible. The computational requirements of the procedure are modest, which should enhance the application of the constrained shortest path problem.