Optimization of Goods Distribution System Using Variable Neighborhood Search Method

Abstract

The problem of distribution of goods in the field of logistics is one of the problems that often occur today. In the distribution process, all companies expect to be able to optimize distribution costs to be more efficient. This makes the capacity of the transportation and distribution system must also be increased in order to create an efficient logistics system. To achieve this goal, producers must develop a proper distribution plan, because wrong distribution can lead to unsatisfactory distribution. Vehicle Routing Problem with Time Windows is a mathematical problem to determine the route for several identical vehicles in serving a number of customers with a certain stone limit. VRPTW is a Nondeterministic Polynomial Time problem, so the exact optimization method is difficult to solve the case. The purpose of this research is to solve the problem of VRPTW with the objective function that is the optimal distance by considering the time limit and vehicle capacity. This research uses the Variable Neighborhood Search algorithm in the solution search stage to solve the problem. Relying on the principle of selecting the nearest customer to be added to the vehicle route, the VNS method can utilize the concept of a varied solution environment, which enables the algorithm to explore the solution space in a more flexible way. The results of research on 100 customers obtained a total average mileage of 624.49. The total mileage without using the VNS method is 888.69, so the difference is 264.2 or there is a mileage efficiency of 29.73%.