Tuning of the Particle Swarm Optimization algorithm parameters in solving the Open Vehicle Routing Problem using the Data Envelopment Analysis

In this research an alternative approach is proposed for tuning of the algorithm parameters instead of the conventional empirical analysis. Since in using the meta-heuristic algorithms the quality of solution is directly related to the parameters setting, so they should be tuned before solving the large scale problems. Usually setting of the algorithm parameters spends a lot of time as well as designing the operators in the algorithm, so proper approaches should be used for tuning of them. In this research, according to the importance of the parameters setting of meta-heuristic algorithms, an approach is applied for tuning of parameters in solving of the OVRP by the Particle Swarm Optimization algorithm.. The proposed approach for the tuning of parameters is based on a nonparametric method which is Data Envelopment Analysis (DEA). Examples of numerical simulations are used in this study for analysis and comparison. To investigate the validity of the proposed method in tuning of algorithm parameters, this method is compared with the Taguchi method. The results show that the DEA approach is a proper one to set the PSO algorithm parameters in solving of the OVRP and leads to act the algorithm more efficient.