Reactive power distribution network optimization neural network based on particle swarm optimization

摘要

Particle swarm optimization algorithm instead of back-propagation algorithm to train the neural network weights and threshold training results from the two algorithms can be seen that PSO-based neural network has a faster convergence rate, and can be used to overcome the single particle swarm optimization algorithm is difficult to achieve the desired results and the single use of the phenomenon of BP algorithm is easy to fall into local optimal solution of the defect phenomenon. Using particle swarm optimization trained neural network to optimize the reactive power compensation devices, the experimental results show that the optimization effect is obvious, is an effective and practical method for reactive power optimization.