MODEL AND DISCRETE PARTICLE SWARM OPTIMIZATION ALGORITHM FOR VEHICLE ROUTING PROBLEM WITH MULTIPLE TRIPS, TIME WINDOWS, AND BACKHAULS

• Main Author: Topani Lubis, Alek
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/36823
• Institution: Institut Teknologi Bandung

Vehicle routing problem is a study to determine the optimal vehicle route to obtain minimal transportation costs. These studies have been conducted many times to find a solution of the the problem based on real systems. However, not all can be completed. This is due to the complexity of the problems. One of the problem’s characteristics is the process of chemical products pickup and delivery. In this case, the process of pickup and delivery must be done separately because of the danger that potentially brought out by chemical reactions. Another problem is the depot and the customers have service time windows. As a result, service must be carried out only at certain time. In addition, to reduce the fixed costs, the vehicles are allowed to carry out multi trips during the planning horizon. A mathematical model is developed to solve these problems. The objective function is to minimize the vehicle used and the total duration of the tour. The solution model is solved by the Branch and Bound (BNB) algorithm with the help of LINGO 17. In addition, the metaheuristic algorithm is also developed to tackling the computational time problem in the analytical method. The DPSO algorithm is used as its metaheuristic algorithm. In this study, the mathematical model of MRKRMJWB as well as the algorithm were successfully developed. The result is that the DPSO algorithm was able to reach the optimal solution with a difference in analytical solutions of 0.03% with a comparison of computing time of 82.38%. In addition, the algorithm was also tested with large scale data. The results of the algorithm succeeded in completing it with a variance coefficient of 6.1% for data 1 and 7.5% for data 2 with five replications.