Solving Capacitated Vehicle Routing Problem (CVRP) for the Environments with Circular Obstacles

Abstract

In capacitated vehicle routing problem (CVRP), a fleet of vehicles with a certain capacity starts at a central depot and returns to this starting point after serving some customers by using the optimal set of routes with the minimum cost. However, in real-life, environments may include obstacles such as holes, machines, or trees of different sizes. In this research, a CVRP extension is proposed that is the problem of the classical one for the environments with various sized circular obstacles. To solve this problem, a hybrid meta-heuristic algorithm based on genetic algorithms improved by a local search was developed. Additionally, a visual simulation tool was designed to place obstacles and locations in the working space. The developed algorithm was tested for different customer-obstacle counts and obstacle sizes with various obstacle occupancies in the environment. The results obtained are presented and the problem’s potential applications are discussed.