Many-objective Reactive Power Optimization Using Particle Swarm Optimization Algorithm Based on Pareto Entropy

摘要

This paper presents a many-objective reactive power optimization model which consists of minimum active power loss, minimum node voltage deviation, maximum static voltage stability and maximum power supply capability. To efficiently solve this model, a novel approach by using particle swarm optimization is proposed. This approach is called many-objective particle swarm optimization algorithm based on Pareto entropy which adopts loose Pareto dominant relationship and maps the Pareto front from cartesian coordinate system to parallel cell coordinate system, thus designing evolutionary strategies using Pareto front's distribution entropy and entropy difference in the new coordinate system. The presented algorithm is capable to balance convergence and diversity of the approximate Pareto front. Moreover, cell dominant intensity and individual density are introduced to assess the individual environment fitness of the Pareto optimal solution, and we hereby design the selection strategy of the global best solution. Simulations based on the IEEE 14-bus systems demonstrate the effectiveness of the proposed model and the efficiency of the proposed algorithm.