Abstract
Electric Power System (EPS) has the characteristics of high dimension, nonlinearity, strong coupling and uncertainty, which makes many problems in intelligent control of EPS not be properly solved. GA (Genetic algorithm) is a search algorithm based on natural selection and genetic mechanism, which is suitable for solving reactive power optimization problems in EPSs. In order to realize intelligent control of EPS, this article proposes a reactive power optimization model of EPS based on improved GA. In the stage of evolution, the crossover probability and mutation probability are dynamically adjusted according to the actual situation of the population. The simulation results show that the accuracy of the improved GA can reach 95.21%, which is about 18% higher than the traditional neural network algorithm and ID3 algorithm. Therefore, the algorithm ensures the convergence of GA and improves the optimization ability of the algorithm. Its efficiency and accuracy are excellent, which can provide theoretical and technical support for intelligent control of EPS.