Application of Genetic Algorithms to Solve MTSP Problems with Priority (Case Study at the Jakarta Street Lighting Service)

Abstract

Transportation is one thing that is very important and is the highest cost in the supply chain. One way to reduce these costs is to optimize vehicle routes. The Multiple Traveling Salesman Problem (MTSP) and Capacitated Vehicle Routing Problem (CVRP) are models that have been extensively researched to optimize vehicle routes. In its development based on actual events in the real world, some priorities must be visited first in optimizing vehicle routes. Several studies on MTSP and CVRP models have been conducted with exact solutions and algorithms. In a real case in the Jakarta City Street Lighting Section, the problem of determining the route in three shifts is a crucial problem that must be resolved to increase worker productivity to improve services. Services in MCB (Miniature Circuit Breaker) installation and maintenance activities for general street lights and priority is given to light points that require replacement. Because, in this case, the delivery capacity is not taken into account, the priority of the lights visited is random, and the number of street light points is enormous, in this study, we use the MTSP method with priority and solve by a genetic algorithm assisted by the nearest neighbor algorithm. From the resolution of this problem, it was found that the travel time reduction was 32 % for shift 1, 24 % for shift 2, and 23 % for shift 3. Of course, this time reduction will impact worker productivity so that MCB installation can be done faster for all lights and replace a dead lamp.