An improved particle swarm optimization algorithm based on simulated annealing
Huafen Yang, Yang You, Zuyuan Yang, Lihui Zhang · 2014
This paper proposes an improved particle swarm-based-simulated annealing method by combine simulate annealing algorithm and swarm particle optimization. An improved annealing schedule is introduced to enhance the performance of particle swarm optimization. The cooling rate is higher at the beginning than at the end of the search process. In this way, the algorithm can explore for solutions in more paths, increasing the probability that the global optima is found. At the same time, particle swarm-based-simulated annealing method introduces the SA metropolis acceptance rule. The metropolis determines whether to accept the new position or recalculate another candidate position according to the fitness function difference between the new and old positions. This enables the solution to jump out of local optimal value, and the vibration is decreased when the searching process is near the end. Experiment results and comparisons with the standard PSO and SA show that the IPSO-B-SA can effectively enhance the searching efficiency and greatly improve the searching quality.