Applying Monte Carlo Techniques to the Capacitated Vehicle Routing Problem

This paper describes a new method for solving the Capacitated Vehicle Routing Problem (CVRP). The CVRP is defined as the problem of delivering goods from a depot to a number of customers with a fleet of capacitated vehicles, minimizing the cost of distributing the goods while respecting the capacity constraints of the vehicles. We have developed a method based on Monte Carlo Simulation (MCS) and the Clarke & Wright Savings (CWS) heuristic that, with similar memory and CPU usage, outperforms previous best known Monte-Carlo-based algorithms for the CVRP. Performance is comparable with that of the current state-of-the-art methods based on metaheuristics.