Comparative Analysis of New Solutions for the Capacitated Vehicle Routing Problem Against CVRPLIB Benchmark

The capacitated vehicle routing (CVRP) is a combinatorial optimization problem (COP) which is belong to NP-hard problem category. This problem includes determining the best option for low-cargo vehicles to serve a number of clients. Every client is required to make a single journey, and the overall demand on each route cannot be greater than the vehicle's capacity. Numerous precise and heuristic approaches have been put out in the literature to address this problem. Specific methods such as branch-binding and branch-pruning assure the search for optimal solutions but are often impractical for large data sets due to their computational complexity heuristic and metaheuristic approaches like genetic algorithms (GAs), ant colony optimization (ACO), simulated annealing (SA), and tabu search (TS), provide optimal solutions in real time, making them more appropriate for bigger instances. Despite these advancements, most studies have not achieved new results compared to those in the standard CVRPLIB benchmark instances. Therefore, this study focused on exploring new results obtained by state-of-the-art approaches in the literature that have solved CVRP. In particular, our aim is to explore new algorithmic improvements designed to improve CVRP regarding to computational efficiency and quality of the solution.