Comparison of Simulated Annealing, Nearest Neighbour, and Tabu Search Methods to Solve Vehicle Routing Problems

Abstract

The high cost of logistics is often caused by the determination of distribution routes, vehicle selection, and vehicle scheduling which is often referred to as Vehicle Routing Problem (VRP). VRP can be interpreted as determining the route with minimal costs to deliver goods from a depot to several different customers with different requests. PT Polar Ice Crystal Semarang is a company producing crystal ice that distributes crystal ice in the Semarang area. PT Polar Ice Crystal Semarang does not have a structured and systematic method in determining distribution routes so the company cannot calculate or know how effective the route travelled by the fleet. So far, the determination of distribution routes is left entirely to the driver of the fleet without any route instructions from the company. Therefore we need a method that can determine the distribution route so that the route passed can be effective in terms of cost and time of distribution. In this research, the determination of the optimal distribution route uses three different methods, namely the Simulated Annealing Algorithm, Tabu Search, and Nearest Neighbour. Then the best method is chosen that can produce the optimal route for the company. Based on the calculation results, it can be seen that the optimal route is obtained using the simulated annealing algorithm with estimated travel costs incurred in the amount of Rp. 293,000.17, with a distance of 88.66 km and a travel time of 9 hours 56 minutes using one large vehicle.