Evaluating the Developed and Mutated Particle Swarm Algorithm (DMPSO) for Optimal Reservoir Operation

The increasing in consuming water with good quality and quantity has prompted engineers and designers to plan and present the advanced plans for optimal utilization of water resource systems. Heuristic optimization methods such as particle swarm algorithm that is investigated in this paper are among the latest methods that have been widely used in recent decades. In this study, a developed model of particle swarm algorithm was described and the particle swarm algorithm was evaluated with a nonlinear and constrained single-reservoir optimization equation. For this purpose, at the first step the system was optimized for a period of five years. Lingo software, a non-linear programming model, was used to compare the results. The value of the objective function in the developed algorithm showed only 0.12 differences with the global optimum. After analyzing the sensitivity of the parameters and the efficiency of the algorithm in the five-year period, the ten-year period was selected for optimization. In this case, the objective function had less than 1% difference with the global optimum as well. In both cases, the developed particle swarm algorithm gave better results than the particle swarm algorithm, and sticking in local optimizations value could be removed. Therefore, it could be concluded that the developed particle swarm algorithm had a high ability to solve the complex problems of water resources systems operation.