Givsy Mea Cendani, Sapti Wahyuningsih, Mohamad Yasin
Mathematics is one of the sciences whose application can be used in solving distribution problems. The study of graph theory, especially on the Vehicle Routing Problem with Simultaneous Delivery and Pickup (VRPSDP) material, has a special problem, namely that demand and return of goods are carried out simultaneously. The focus of this article is the study of the VNS-GA Hybrid algorithm used in solving VRPSDP. The main stage consists of two stages, namely the initial stage of forming a solution using the Genetic Algorithm and the stage of improving the solution using the Variable Neighborhood Search (VNS) Algorithm. The implementation of the algorithm uses the Borland Delphi 7.0 programming language. Program inputs are vehicle capacity, population size (popsize), crossover probability (Pc), mutation probability (Pm), number of iterations, maximum number of iterations of the disturbance phase. The output produced by the program is the route formed along with the total distance and graphic visualization. The program has been tested using 8, 20, 40, 75 customers and a standard dataset, namely 100 customers (c101 and r101) and 200 customers (c1_2_1 and r1_2_1). Among the standardized datasets, the smallest gap value is dataset c101 with a gap value of 2.8 % from the Tabu Search algorithm solution. In this case, the VNS-GA Hybrid algorithm approaches the optimal results using a standard dataset. © 2024 Author(s).
Departement of Mathematics, Mathematics and Natural Science Faculty, Universitas Negeri Malang, Malang, Indonesia