A Probabilistic Hill-Climbing Algorithm for the Single-Source Transportation Problem

This paper proposes a probabilistic hill-climbing algorithm, called PH, for the single-source transportation problem (STP). PH is tree search algorithm in which each node contains an assignment (AP) transformed from STP being solved. The transformation converts source’s product units into lots; lot equals multiple units. AP aims to find optimal of lots destinations minimize total cost. uses Hungarian method solution every node, STP. For root (as initial current node), number set be small enough guarantee generation feasible To generate subsequent level, branched child nodes, nodes sources modified by adding one more specific different source. Consequently, provides that better than or same as node’s solution; however, some nodes’ solutions may infeasible due insufficiency capacity. If all cannot solution, stops its procedure. diversify search, selects new way, instead always selecting best node. experiment’s results this reveal performance three variants PH.