SUMMARY
This research studies the distribution network redesign of an actual electronics company. The problems are formulated based on multi-echelon capacitated Location Routing Problem (LRP) with two commodities: home products and service items. The objective function consists of three components: facility cost, closing cost of facility and transportation cost. We propose solution method based on clustering technique. The problem is decomposed into the Facility Location Allocation Problem (FLAP) and the Multi-Depot Vehicle Routing Problem (MDVRP). MDVRP is solved by clustering method and feed the results to the modified FLAP to allocate the demand nodes to facilities and configure all distribution networks, for the 2nd and 3rd echelon. The distribution is divided into five region zones. Previously, each region was operated independently but this research compares the solutions from solving each region independently and solving all five zones simultaneously. The results indicate that the proposed solution method can achieve computation time and total cost that are comparable to ones obtained from solving the problem to optimality. Exact approach can only solve small and medium problems, whereas the proposed solution method provides the acceptable solution of real-life largest problem in limit of computation time. Finally, we perform sensitivity analysis on the results.