Random Spanning Tree Based Improved GA for Distribution Reconfiguration

Using traditional genetic algorithm (GA) to solve distribution network reconfiguration, the required radial network structure can not be ensured and a large number of infeasible solutions are brought about. Although some improved methods were put forward, they either are of computational complexity or can not completely settle the problem. In this paper, the strategy of searching randomly spanning trees is introduced to form an improved GA. The method proposed in this paper is simple and convenient to solve distribution network reconfiguration. Combined with graph theory, this algorithm can ensure any reconfiguration scheme characterized with radial structure of network. Numerical tests on both 33-bus and 69-bus networks show effectiveness and advantage of the proposed algorithm.