Extended Pruning-Grafting operators for efficient distribution network reconfiguration

Network reconfiguration is a complicated combinatorial, non-differentiable, constrained optimization problem due to the incorporation of many candidate switching combinations. Introduction of evolutionary algorithms (EAs) for application to distribution network operation has opened many new opportunities. However, these methods suffer from high computational burden. Performance of the EAs is significantly affected by modeling of the problem and employed operators. This paper employs a branch based modeling of distribution network and proposes two new EA operators which are an extension and redefinition of PAO and CAO operators to be utilized for a more efficient application of EAs to the network reconfiguration problem. Performance of the new operators has been compared to original operators as well as an enhanced version of conventional crossover/mutation operators. Simulation results have proved efficiency of the proposed method.