A Novel Hybrid Multiagent-based Particle Swarm Optimization for Distribution Network Reconfiguration

Qianjin Liu,Chuanjian Li

DOI: https://doi.org/10.1109/APPEEC.2010.5448918

2010-01-01

Abstract:A novel method is proposed to solve the distribution network reconfiguration. The proposed method combines the binary particle swarm optimization (BPSO) with discrete particle swarm optimization (DPSO) and multi-agent system (MAS). It also incorporates self-learning mechanism and the survival of the fittest which is a part of natural selection for reconfigure distribution network. In the algorithm, each particle is thought to be an agent and all agents live in a lattice-like environment. The competition and cooperation operator is preformed on every agent. If the agent is a loser, a new agent will occupy its lattice-point which is determined by BPSO and DPSO. If the agent is a winner, it can perform the self-learning mechanism, so as to obtain the better fitness value. Finally the next iteration particles are determined by the fitness value of previous particles, according to the survival of the fittest mechanism. The proposed method applied for reconfiguration distribution network is evaluated on a typical example of PG&E 69 nodes distribution system and compares with other methods. The result shows that the method has superior features, including good computation efficiency, good convergence characteristics, and high-quality solutions etc.