Decision Support System for Optimizing Rastra Distribution Routes Using Genetic Algorithm

Abstract

Vehicle Routing Problem (VRP) is an optimal route design from a group of vehicles that deliver goods to a set of customers with a certain demand. VRP was widely studied as part of solving the distribution efficiency which minimizes the cost of traveled vehicle. Bulog Subdivre South Surabaya distribute Rastra to every village which has constraint of Multi depot and Split delivery (MDSDVRP). This study aims to minimize the traveled distance of MDSDVRP (Rastra distribution) using  Genetic Algorithm (GA) and to find out the efficiency of the route solution. The research covers the steps to solve MDSDVRP using GA to generate feasible and efficient solution route. Then development of a Decision Support System (DSS) that applies the algorithm is implemented on web platform and the result of route solution is presented on the mobile platform. The system testing is carried out to test the user satisfaction (83.8%) which found that overall users were considered very agree, good, like for each component of user satisfaction. The traveled distance is compared between GA route and routes of the original data from 2013-2017. The efficiency of GA was evaluated and found that the traveled distance from the previous route is reduced by 3.7% (444 km) and in 2017 is reduced the distance traveled by 9.5 % (1,093 km). The GA can generate a better solution and optimize the distance than the original route.