Assessment of Heuristic Algorithms for Solving Real Capacitated Vehicle Routing Problems by Analytic Hierarchy Process

In solving the real-world transport tasks, it is not easy to set up the real problem within the frames of mathematically defined vehicle routing problems (VRP). There are also difficulties in selecting the algorithm that could give the optimal solution of the classified problem. Using the real vehicle routing problem six constructive algorithms were developed and for each algorithm five optimizations i.e. local searches were performed. All calculations were made by using the interactive software environment Venera, and the algorithms were written in the programming language Mars. The solutions have been assessed and inter-compared by means of analytic hierarchy process. For the same problem, the location of the depot in relation to the user cluster was changed. Six mentioned algorithms were performed on the newly created problem, and the newly obtained solutions were analyzed in relation to the previous location of the depot. The following factors were taken into consideration: the number of vehicles, the time necessary to calculate the route and the overall traveled distance of the vehicle fleet. It has been determined that there is dependence of the solutions obtained by certain algorithms regarding the location of the depot with regard to the customers as well as the possibility of achieving savings by changing the location of the depot.