Abstract: Particle swarm optimization is a new evolutionary computation method of the group and is widely used in a few projects. Because the convergence speed is slow and easy to fall into local minimum, the paper proposes a hybrid particle swarm optimization algorithm based on differential and simulated annealing. Through analysis and combining of their respective advantages of the three kinds of evolutionary algorithm, an improved particle swarm optimization is obtained.

Key words: particle swarm, difference algorithm, simulated annealing, optimization