Improved Population Migration Algorithm And Application

Population Migration Algorithm(PMA)is a novel method to search global optimal value based on the behavior of population migration proposed by Zhou Yong-hua and Mao Zong-yuan in 2003. This algorithm has the certain global search ability. However it has such disadvantages of local optimization and slow convergence speed. For overcoming these shortcomings, this paper proposed an improved PMA which adopted norm and could keep more than two individuals. Utilizing the improved PMA to train Artificial Neural Network (ANN), we established a new Artificial Neural Network model PMA ANN. PMA ANN had a higher ability of searching for globally optimal solutions and a quicker convergence speed comparing with the Genetic Algorithm,Artificial Fish-swarm Algorithm and BP Algorithm. In the end, a PMA ANN forecast model was established by applying PMA ANN to the electrical power system short-term load forecasting, and the tests indicated that it had improved forecast precision.