MODEL AND THE TWO LEVEL NEIGHBOURHOOD SEARCH ALGORITHM FOR VEHICLE ROUTING PROBLEM WITH BACKHAULS AND HETEROGENEOUS VEHICLES IN CATERING BUSINESS

• Main Author: Ayu Chumaira, Soraya
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/54134
• Institution: Institut Teknologi Bandung

Managing routings is one of the essential components in the catering business. Minimizing the cost of transportation will be obtained from minimizing distance by finding the optimal route. This problem is called the vehicle routing problem. This research presents vehicle routing problem variants, a VRP with backhauls, a single depot, heterogeneous vehicles, and a single trip for each vehicle (VRPB). This variant of VRP adapted from catering business behavior that prioritized deliveries than pickups. The mathematical model can not find the solution in 5 hours. An algorithm with a two-level neighburhood search has been developed to solve the problem. For the effectiveness of the algorithm, it finds total distances less than the original routing by the caterer. It reduces 7.88 percent of the total distance. For the efficiency, it provides a feasible solution in average computation time is 182,59 s. The algorithm is tested using small and big problems for datasets 2 with a constant composition of vehicles for all instances and dataset 3 with the various compositions of vehicles for all instances