An Improved Neural Network Method for Solving Economic Load Dispatch in Power Systems

摘要

This paper presents a novel optimization approach combining feed forward neural network based particle swarm optimization (PSO) and the lambda-iteration method to solve the economic load dispatch problem in electrical power system (ELD). The feed forward neural network trained by the PSO instead of conventional back-propagation algorithm is developed, which can speed up convergence and overcome local minima. Simulation results have demonstrated that the proposed approach is feasible and effective for the application in ELD.