Memetic Heuristic Approach for Solving Truck and Trailer Routing Problems with Stochastic Demands and Time Windows

Abstract

Manufacturers and service providers often encounter stochastic demand scenarios. Researchers have, thus far, considered that the deterministic truck and trailer routing problem (TTRP) cannot address the prevailing demand uncertainties and/or other complexities. The purpose of this study is to expand the deterministic TTRP model by introducing stochastic demand and time window constraints to bring the TTRP model closer to reality and solve the model in a reasonable timeframe by administering the memetic algorithm (MA). This paper presents a model that can be applied in stochastic demand conditions. To employ the MA, various crossovers, mutations and local search approaches were applied. First, two experimental tests were carried out to show the validity and consistency of the MA for solving stochastic TTRPs. The results can be compared with the VRPSDTW (vehicle routing problem with stochastic demands (VRSPD) with time windows) solution obtained using large neighbourhood search (LNS) of earlier research. The average results from Tests 1 and 2 achieved by MA are 514.927 and 516.451. However, the average result obtained by LNS is 516.877. Therefore, the MA can generate results. Thus, MA is found to be suitable for solving truck and trailer routing problem(s) under stochastic demand with a time window (TTRPSDTW). Moreover, 54 benchmark instances were modified for this case and the initial feasible solutions were generated for this purpose. The solutions were significantly improved by the MA. Also, the problems were tested using sensitivity analysis to understand the effects of the parameters and to make a comparison between the best results obtained by MA and sensitivity analysis. Since the differences between the results are small, the MA was found to be appropriate and better for solving TTRPSDTW. The paper also gives some suggestions for further research.