摘要：AbstractThis paper presents a new metaheuristic method for distribution network reconfigurations. In recent years, distribution network reconfigurations play an important role to reduce the network loss with sectionalizing switches and shunt capacitors. Distribution network reconfigurations may be formulated as a combinatorial optimization of the NP hard problem. In this paper, Parallel Discrete Evolutionary Particle Swarm Optimization (PDEPSO) is proposed to deal with distribution network reconfigurations. PDEPSO is Discrete Particle Swarm Optimization (DEPSO) with the island model of parallel computation in which DEPSO integrates EPSO of adaptive PSO with the sigmoid function to handle combinatorial optimization problems. Furthermore, to improve the performance of PDEPSO, this paper develops a method that sets the particle speed limit of DEPSO to maintain the solution diversity and to evaluate better solutions. The effectiveness of the proposed method is demonstrated in the IEEE 69-node distribution system.
