Abstract: According to the load characteristic of power system and the application limitation of grey model GM(1,1),vector is introduced into the computing formula of background value array. A new model GM (1,1,)which bears higher adaptability is constructed. Particle swarm optimization algorithm is adopted in order to solve the value of a as it has the virtue of optimum-seeking. Then a particle-swarm-optimization-based grey model PSOGM is proposed and practical examples are given. The simulation results indicate that PSOGM model gives better precision and has wider application field.

Key words: power load forecasting, grey model, background value, particle swarm optimization