Improved Differential Evolution for Solving Optimal Reactive Power Flow

Differential evolution (DE) is a promising evolutionary algorithm for numerical optimization. However, with descending population diversity, the DE will also encounter premature convergence as other evolutionary algorithms (EAs). Based on analysis of premature convergence of DE and close observation on various improved schemes of EAs, two simple improved DE schemes are developed in the paper. Comparative simulation on the optimal reactive power flow problems shows that the scheme utilizing the variable population size and population reinitialization technique has the best performance. Parameter setting of the schemes has also been investigated in the paper.