Reactive power optimization based on improved particle swarm optimization algorithm with boundary restriction

At present, the particle swarm optimization algorithm is not effective in dealing with discrete variables, avoiding local optimization, and satisfying all inequality constraints of voltage and power factor. Therefore, this paper employs variable reflection and integralization to find the discrete correspondent of continuous variable in the particle; and introduces chaos strategy to the searching process to strengthen the capability of finding global optimization solution. Then this paper improves the optimization solution by adjusting the voltage and power factor that exceed limits with "nine palaces" strategy, to ensure the particles in feasible solution space. At last, this paper tests the proposed algorithm with the standard IEEE sample system and some actual network planning. The comparison of calculation results with those in other literatures proves that the new algorithm has more advantages at searching speed and quality.