Non-smooth/non-convex economic dispatch through modified particle swarm optimization

Abstract

Economic Dispatch (ED) is the foremost typical task in the power system engineering. Its aim to minimize the total generation cost while satisfying system constraints. Considering some practical features like prohibited operating zones, transmission losses and ramp rate limits ED becomes a non-convex, non-smooth, non-linear, and multi-constraint optimization problem. To solve complex optimization problems, particle swarm optimization (PSO) have proved noticeable success. However, it may suffer to trap at local minima, slow convergence and poor solutions. Thus, to address these issues and solve ED problem, a modified particle swarm optimization (mPSO) is introduced. In mPSO, new factors for inertia, cognitive and social components (i.e. acceleration coefficients and inertia weight) are familiarized. As per the concerts these factors are dynamically changed and properly maintained the diversity of the PSO. The projected mPSO is implemented to solve three altered unit test structures (3, 6 and 15) of ED problem. The experimental and compared results shows that proposed mPSO produce more appropriate economic cost.