An improved distribution system reconfiguration using hybrid GA with PSO

This paper presents an efficient and accurate way of solving radial distribution system reconfiguration (DSR), which plays an important role in distribution automation for realizing smart grids. It deploys different heuristic optimization approaches to resolve the desired optimum configuration and to efficiently reconfigure the connectivity of the distribution networks. The objective is to minimize the system power loss while the voltage in each bus is limited to some allowable range, and the topology of the system is kept radial. In this paper, a hybrid algorithm consisting of particle swarm optimization (PSO) and genetic algorithm (GA) is proposed. The first part of the hybrid approach is based on a modified PSO where the initial swarm of particles fit the radiality constraint and it introduces external randomness to velocities and locations with certain probabilities when particles are in equilibrium or close to equilibrium state. The second part is a modified GA which receives its initial population from best solutions of the modified PSO and uses adaptive mutation for introducing population diversity. In addition, the particle's location in PSO and each chromosome in GA are repaired in such a way that the radiality constraint is always satisfied. The validity and the effectiveness of the proposed method has been tested using the standard IEEE 33-bus distribution network. The results show that the proposed method is robust and delivers a minimal average power loss of independent runs with reduced computational time.