Algoritma Simulated Annealing untuk Optimasi Rute Kendaraan dan Pemindahan Lokasi Sepeda pada Sistem Public Bike Sharing

Abstract

The public bike-sharing system has a problem where the number of bicycles at the docking station needs to be balanced to ensure system user satisfaction. The usual solution is to distribute bicycles so that system users can still park for locations that are usually full of bicycles or pick up bicycles at locations that normally lack bicycles. The purpose of solving this problem is to get a vehicle route with the total operating costs of the vehicle. The full vehicle operating costs are associated with the full time taken by the vehicle to distribute the bicycle. Besides, there are also penalty fees related to the lack of bikes or parking slots at the time of operation of the public bike-sharing facility. In this study, two variations of the simulated annealing (SA) algorithm were developed to solve the SBRP problem called SA_BF and SA_CF. The data used comes from a Velib bike-sharing system case study in Paris, France. The results of the experiment show that both the SA_BF and SA_CF algorithms succeeded in solving SBRP. This algorithm has an average difference of 2.21% and 0.36% of the Arc-Indexed algorithm (AI) from previous studies in the first dataset. As for the second dataset, Tabu Search algorithm, SA_BF and SA_CF obtained an average difference of 0.65%, 1.08% and 0.38% of the optimal results.