Implementation of the ACS-RVND algorithm on the VRP variant and its application to distribution optimization

Abstract

Abstract Distribution optimization has an important role to distribute products to consumers. The selection of the right route will have an impact on the distribution process resulting in minimal costs. The application of graph theory that can be used to determine the optimum route in the distribution process is the study of the Vehicle Routing Problem (VRP) variant. The VRP variants discussed in this article are MDVRP and VRPTW. The method used to solve the VRP variant of this article is the ACS-RVND algorithm. The ACS-RVND algorithm consists of several main stages, namely the initial solution formation stage using the ACS algorithm, the solution improvement stage using the RVND procedure and the acceptance criteria stage. The data needed for the application of the ACS-RVND algorithm in solving the distribution optimization problem of MDVRP cases are the number of depots, the number of customers, the capacity of the vehicle, the number of ants, the number of iterations, the number of customer requests and the distance between customers. While in the case of VRPTW data multiple depots are replaced by time windows depot and service time. The results of solving distribution optimization problems in the form of the route, the total distance traveled and the result of the route in the graph model. The performance of the ACS RVND algorithm can be compared with the performance of the ant colony optimization (ACO) algorithm and the Hybrid Ant Coloy Optimization (HACO) algorithm. Analysis of the results using several dataset test cases showed that the ACS RVND algorithm on MDVRP obtained a better solution than the ACO algorithm and the ACS RVND algorithm on VRPTW was better than the Hybrid Ant Coloy Optimization (HACO) algorithm when viewed from the total distance distribution route.