Restricted evolutionary strategy for optimal capacitor switching in distribution systems

To reduce the computing time and improve the efficiency of traditional evolutionary programming, this paper proposes a restricted evolutionary strategy for the optimal capacitor switching in distribution systems. The reactive power compensation ranges are obtained by the physical constraints of the distribution network, and thus the initialization and evolution are restricted to a reduced optimization space. The algorithm's starting points proposed are arranged around the center of the optimization space to prevent infeasibility of the initial solutions. The smaller population and more generations strategy are used to reduce large numbers of infeasible solutions during the optimizing process. Based on the randomly dynamic-step evolutionary method, which takes different steps during different stages of the optimization, the effectiveness and efficiency of optimization are greatly improved. The strategy proposed is tested by different distribution systems, its high efficiency demonstrating its promising online application.