A genetic algorithm with local search for multi-product inventory routing problem with a fleet of multi-compartment vehicles

Abstract

This paper presents a genetic algorithm (GA) with local search method to determine the solution for the inventory routing problem (IRP) with a homogeneous fleet of multi-compartment vehicles. The objective is to minimize the total cost including the rental cost, the travelling cost, and the inventory cost. Each customer is allowed to have multiple visits by vehicles. The multi-product with known demands and limited tank capacities at customers are in our consideration. The mathematical model for this IRP is presented and classified as the mixed integer programming. Since the IRP is considered as the NP-hard problem, GA is developed to deal with the large-scale problem. The proposed GA with local search method is utilized to determine both the order quantities and routes for distribution. The chromosome representation, GA operators, and GA parameters are described in this paper. The numerical examples reveal that, for the problems having high complexity, the proposed GA can yield better quality of solutions than the solutions of the optimization software namely CPLEX 12.4. Moreover, the computational time of the proposed GA is significantly lower than that of CPLEX 12.4 for the large-size problem.