A many-objective gradient evolution algorithm for solving a green vehicle routing problem with time windows and time dependency for perishable products

Abstract

Abstracts All over the world, huge numbers of perishable goods are transported from suppliers to recipients every day. Some perishable products such as foods and medicine require a special treatment during delivery due to their limited lifetime, and all perishable products must be transported as quickly as possible before they spoil. In addition to the time limitation on transport, high frequency of transportation may lead to high transportation cost, making the optimization of transportation cost for delivering perishable products very important in this industry. In addition, high frequency of transportation may also contribute to air pollution. In order to address this problem, this study proposes a green vehicle routing problem (VRP) for perishable products which optimizes the operational cost, deterioration cost, carbon emissions and customer satisfaction. The proposed VRP model also considers time windows, different travelling time during peak hour and off-peak hour, and working hours. This paper solves the proposed model using a many-objective gradient evolution (MOGE) algorithm. GE algorithm is a new metaheuristic which was originally proposed for continuous problems with a single objective. However, this study improves the original GE algorithm with discretization, non-dominated sorting, and crowding distance approaches. The proposed model and algorithm are employed to solve a fruit distribution problem. The experiment results show that the proposed MOGE algorithm has more promising results than other algorithms.