Distribution Network Reconfiguration Based on Adaptive Bi-Group Particle Swarm Algorithm

This paper presents an efficient algorithm for loss minimization in distribution network reconfiguration. The algorithm divided the initial particle swarm into two different sub-groups which perform simultaneously. So there are more particles with high inertia weights in the prophase of the evolution in order to make the large-scale searching in solution domain easily, and in the anaphase, these particles inertia weights will descend in linearity which improves the ability of local searching. As the changing of the particles quantity of the two sub-groups and the effect of the different inertia weights, it can achieve a good searching performance and overcome the limitations of PSO, which has a slow convergence velocity and easily being relapsed into local extreme. Then on the basis of improved binary particle swarm optimization algorithm and in combination with the breaking circle method, to make the updated particles meet the requirement of the radial network by hundred percent and accelerate the calculation speed. Finally a typical example of the IEEE 33-nodes network was simulated through the proposed algorithm to calculate the network power losses, and the simulation results was compared with those of modified binary PSO algorithms. The comparison results show that the proposed method has a higher searching efficiency and faster convergence rate.