Solving Multi-criteria Vehicle Routing Problem by Parallel Tabu Search on GPU

Transportation plays a crucial role in both production and service industry. In modern times, the importance of supplying the goods on time to warehouses, production units and finally to the customers is not lost on logistic companies. Vehicle Routing Problems (VRP) models evolved to be more advanced, which led to the growth of computational complexity. Optimizing transportation routes for companies means performing complex computations and doing so in the shortest possible amounts of time. Graphics Processing Units (GPUs) provide massive computation when the needed operations are properly parallelized. nVidia GPUs are equipped with Compute Unified Device Architecture (CUDA), so applying parallel algorithms is not limited to complex workstations or specialized computers. This work emphasizes the value of using parallel Tabu Search (TS) algorithm over sequential TS algorithm and its application to multicriteria discrete optimization of Distance-constrained VRP.