Vehicle routing problems (VRPs) have received a lot of attention both from the theoretical and practical perspective. The differences between the mathematical and engineering approaches are considered. A definition is given of VRP. VRP aims at designing a minimum cost set of routes so that each route begins and ends at the depot, the total load on each vehicle is smaller or equal to its capacity, and that each customer is served by exactly one vehicle. VRP heuristics can be divided into constructive methods, 2-phase algorithms and meta-heuristics. The performance criteria for VRP heuristics are normally related to four main parameters: accuracy of the solution reached, computation times, simplicity and flexibility. A simple and flexible heuristic for real-life VRPs is presented: STRIDER (System to Realise Intelligently and Dynamically Economic Routes). The different stages of the iterative method are outlined and the computational results are presented. The limits of the method are described.
Abstract