Optimization of Fresh Food Distribution Route Using Genetic Algorithm with the Best Selection Technique

Abstract

All along the food supply chain, managers face the challenge of making important cost-optimized decisions relating to transportation and storage conditions. An efficient management of food products requires the consideration of their perishable nature to solve the safety problem, and logistics efforts in minimizing the total cost while maintaining the quality of food products above acceptable levels. This paper addresses a methodology to resolve a capacitated model for food supply chain (FSC). The model is a constrained mixed integer nonlinear programming problem (MINLP) that computes the cost of quality internally for a three echelon FSC, tominimize the total cost under overall quality level constraints, due to the perishable nature of the products, and the other constraints of demand, capacity, flow balance, and cost. The practical application of the model is demonstrated using two approaches: an exact method based on the Branch & Bound technique, and a Genetic Algorithm solution method. Then, we propose a comparison of different GA selection strategies, such as Tournament, Stochastic Sampling without Replacement, and Boltzmann tournament selection; the performance of each selection method is studied using paired "T-Test" of statistical analysis. The results of computational testing are reported and discussed, and implications and insights for managers are provided by studying instances of practical and realistic size. It was evident that the tournament selection is more likely to produce a better performance than the other selection strategies.