Optimization of Very Large Scale Capacitated Vehicle Routing Problems

Abstract

Transportation activity is one of the main contributors to the overall logistics costs of a company. One of the ways to minimize the cost that comes from this activity is through the optimization of the distribution route. However, with the growth of delivery points, the possible routes combinations are growing in a nonlinear fashion, increasing the complexity to obtain the optimal combination of the distribution routes. In this paper, we present a simple modified clarke-wright algorithm with the implementation of heuristic pruning to deal with time and space complexity in a case of very large scale capacitated vehicle routing problem. The resulting algorithm can solve a very large routing problem with several thousand customers within a minute and produce relatively good solutions compared to the state-of-the-art algorithm, especially for the problem with a depot located at the center. Furthermore, the algorithm is also tested to solve a very large scale problem of FMCG company in Indonesia with 32,223 delivery points.