A new algorithm based on non-dominated sorting differential evolution for multi-objective optimal load dispatch

A multi-objective optimal load dispatch model is proposed for minimizing both total purchase cost and emissions considering the difference of power purchase price and emissions between each power plant. And a new NSDE algorithm, which is organic integration of Pareto non-dominant sorting and differential evolution algorithm and improves the crowding distance mechanism and mutation strategic in evolution effectively, is designed for the model. Simulation on a power system with 6 generating unit shows that the algorithm can overcome disadvantages in conventional multi-objective algorithms with apparent improvement on search speed, integrality of Pareto front, distribution of non-dominant solutions and convergence and can achieve more accurate and diversiform Pareto non-dominant optimal set.