Models and Algorithms for Multiagent Hierarchical Routing with Time Windows

Abstract

The problem of modeling real logistics systems arranged in a hierarchical manner is considered. Clusters of lower level consumers are formed that meet the time window (TW) constraints for each consumer and the cluster as a whole. In each such cluster, a traveling salesman’s route is constructed and the vertex closest to the central node, which is the vertex of reloading goods from heavy vehicles (Vs) to light Vs serving consumer clusters, is selected. The transshipment vertices, in turn, are combined into higher level traveling salesmen’s routes, taking into account TWs for routes of this level. The software implementation is tested on well-known networks. The technique is applicable for the synthesis of the central distribution center and system distribution centers of the lower level, as well as for calculating the required number of vehicles (agents).