Performance comparison of GA and DEA in solving distribution system reconfiguration problem

Abstract

Distribution network reconfiguration (DNR) is a crucial issue in the operating horizon of distribution systems. In distribution networks tie switches will be changed to obtain an appropriate network configuration with minimum losses. Distribution networks have hundreds of switches and determining the best status of switches is a complicated combinatorial, non-differentiable constrained optimization problem, which can be solved using heuristic optimization algorithms. Many researches have been focused on introducing appropriate search algorithms to this real-time optimization problem. To judge which method is best suited for this particular problem, performance comparison is necessary. In this paper GA and DEA are compared solving reconfiguration problem from convergence rate and computational time point of view.