A two-phased approach to the supply network reconfiguration problem

This study focuses on a supply network problem where resource inputs are constrained to achieve performance goals for the reconfigured distribution system. Simulated annealing is applied for solving this reconfiguration problem. The supply network is comprised of distribution centers that are geographically distributed and have differing operating and maintenance cost structures on distribution-related resources. The distribution planning methodology employed in the heuristics has two phases: Phase one incorporates a strategic-based decision making step in which the ‘best’ set of distribution centers are selected open; Phase two incorporates an operationally-based decision making step in which the heuristics perturb the customer assignments and resource assignments. The heuristics are tested on a set of 450 problems of various sizes and structures. Computational results suggest that the simulated annealing heuristics generate near optimal solutions quickly and are well suited for evaluating supply network reconfiguration. Computational results also suggest that enhanced annealing heuristics proposed in this paper are better than the standard annealing approach. The supply network problem structure was exploited to achieve very good solution results.