Integrated mutation strategies in a biologically inspired algorithm to solve the vehicle routing problem with moving shipments at the cross-docking centre

Abstract

The cross-docking is an innovative logistic technique in a supply chain. In a previous study, some internal operations at the cross-docking centre were integrated to develop a model for the Vehicle Routing Problem with Moving Shipments at the Cross-Docking Centre (VRPCD-MS). However, only the small-scale instances were used to obtain the exact-optimal solution and the study recommended applying meta-heuristic approaches to reach near-optimal solutions to the large-scale instances of VRPCD-MS. This study attempts to employ Genetic Algorithm (GA) with multiple-mutation strategies ‘SWAP’ and ‘REVERSION’ to solve VRPCD-MS. Initially, GA with SWAP and GA with REVERSION mutation strategies are employed sequentially as two different approaches. Consequently, the solutions of the benchmark problems reveal that there is no strong evidence to recommend any of those two approaches is better than the other. Furthermore, the two-sample t-test concludes that, there is no significant difference in solutions by applying these two mutation strategies alone. It can be concluded that, up to 30% improvement in the solution can be obtained by using GA with merged SWAP and REVERSION mutation strategies together. It further concludes that, on an average, nearly 15% improvement in the solution is reached by the integrated mutation strategies in the proposed GA approach.

Keywords: Cross-docking, Genetic Algorithm, REVERSION mutation, SWAP mutation