Evaluation of Genetic and Particle Swarm Optimization Algorithms Based on Non-Dominating Sorting Approach for Multi Objective Optimization Operation of Reservoirs

Abstract Background and objectives As a crucial issue in aqua sciences, optimizing dam reservoirs exploitation has been studied with a variety of optimization techniques. In recent years, a large number of multi-objective Meta-Heuristic algorithms have been introduced. One of these algorithms is the second version of the multi-objective genetic algorithm with nun-dominated sorting, which was introduced in 2002 by Deb and et.al. In this research, the innovation and aim are to use the particle swarm algorithm with nun-dominated sorting approach and evaluate the efficiency of this algorithm in the optimization of operation of the reservoir performance. Finally, the results are compared with the NSGA-II algorithm, which ultimately leads to a sustainable management policy for water resource systems, and in particular the exploitation of the reservoir. Materials and methods In this research, the multi-objective version of the genetic algorithm and particle swarm optimization were investigated using concepts such as Non-dominated sorting and crowding distance and used to solve the optimization problem of the Mulla-Sadra reservoir in Fars province. The problem of optimization defined with two goals. One of them was minimizing difference between agriculture demand end releases. The second of objective function was maximizing flood storage volume. The two algorithms compared with criteria such as the run time, the number of solutions placed on the Pareto front, standard deviation and performance criteria (dispersion criteria). Results The results of the research indicated that both algorithms have the ability to solve this optimization problem. Also the results indicated that the algorithms has somewhat more performance than some other criteria. The results of the investigation of the runtime of each of the algorithms showed that the performance of the multi-objective particle swarm algorithm (NSPSO-II) is far more than the NSGA-II algorithm, so that the average runtime of the NSGA-II algorithm in the population of 50 The value of 21.3879 seconds is approximately three times the average runtime in the NSPSO-II algorithm with a value of 6.3169 seconds. Regarding the performance criterion, the NSPSO-II algorithm has a better performance than the NSGA-II algorithm. On the other hand, according to the number of solutions on the Pareto front, the NSGA-II algorithm found a lot more solutions on the Pareto front, which is why standard deviation in the NSGA-II algorithm was less than NSPSO-II. Conclusion The NSGA-II algorithm found a lot more solutions on the Pareto optimal front, and the solutions on the Pareto optimal fronted properly covered the Pareto front, unlike the NSPSO algorithm. Also, comparing the solutions in the Pareto optimal front showed that the NSPSO-II algorithm was stepped up to maximize the second target function, while the NSGA-II algorithm moved in the direction of minimizing the objective function. Keyword: Crowding Distance, Dominate, Flood, Malla-Sadra Dam, Performance Criteria.