Research of Distribution Network Reactive Power Optimization Based on Improved Cloud ParticleSwarm.pdfVIP

Research of Distribution Network Reactive Power Optimization Based on Improved Cloud Particle Swarm Optimization BP Neural Network Hong-sheng SU ， Pei-jiong ZHANG Abstracts: In view of cloud particle swarm optimization is convergence precision were not accuracy. Conventional adaptive easily trapped in local minimum and existed slowly convergence particle swarm optimization use the strategies which cant reflect in the distribution network reactive power optimization, the cloud the process of actual optimization search of linear inertia weight particle swarm optimization is improved based on cloud digital features (Ex, En, He), local search and global search which are reduction. Cloud particle swarm based on cloud digital features based on the solution space transform are combined, the random (Ex, En, He) has stability and randomness [12-13]. Cloud particle is crossed by cloud generator of X condition ，the cross generator of X condition can adjust nonlinear dynamically inertia operation of particle according to cloud genetic, variation weight of particles [14-15], but it is trapped in a local minimum operation of the particle according to the normal cloud particle. and less convergence precision in reactive power optimization. In apparent improvement are got in time and storage by using the algorithm, in order to enhance convergence accelerate, train BP this paper, BP neural net