Search

This paper presents an improved Particle Swarm Optimization algorithm (IPSO) for solving the flexible load dispatch problem (FLD) which relates to renewable energy power generation and three types of load composed of electric vehicle (EV), ice-storage central air conditioning and smart household appliance. The objective of the FLD problem is to utilize aggregators of the three types of load to deal with the unpredictability and intermittence of the renewable energy power generation. Facing premature convergence of PSO, the proposed method divides the search process into two stages. In the first stage, every particle selects one particle randomly to learn. The second stage is the conventional PSO algorithm. The total length the optimal particle flights for, the average velocity and the distance of every iteration are used to evaluate performance of PSO and IPSO. The results of a 12-unit system show that the IPSO is more stable and effective than PSO.