Research on adaptive niche differential evolution algorithm for reactive power optimization

Reactive power optimization which is the foundation for optimal control of voltage and reactive power is a very complex non-linear mixed integer programming problem. To deal with this optimization problem, an Adaptive Niche Differential Evolution algorithm (ANDE) is presented to avoid the premature phenomenon in DE. To maintain the population diversity, niche-sharing mechanism is adopted. To improve the rate of convergence, penalty function is used by niche-eliminating mechanism to deal with some low fitness individuals. To improve searching capability of algorithm, niching radius can also be adjusted adaptively on the basis of the relative distance between individuals which reflect the aggregation of population. Using the above method, the global searching ability of this algorithm could be improved. The proposed algorithm has been tested on IEEE-6, IEEE-30 and IEEE-118 bus systems and the calculation proves the effectiveness and practicability of the above algorithm.