抽象的

Multi-strategy multi-agent simulated annealing algorithm based on particle swarm optimization algorithm

Changying Wang, Ming Lin, Yiwen Zhong

Multi-agent simulated annealing (MSA) algorithm based on particle swarm optimization (PSO) is a population-based SA algorithm, which uses the velocity and position update equations of PSO algorithm for candidate solution generation. MSA algorithm can achieve significantly better intensification ability by taking advantage of the learning ability from PSO algorithm; meanwhile Metropolis acceptance criterion is efficient to keep MSA from local minima. Taking into account that different problems may require different parameters for MSA to achieve good performance, this paper proposes a multistrategy MSA (MMSA) algorithm. In MMSA algorithm, three parameter control strategies, multiple perturbation equations, variant number of perturbed dimensions and declining population size, are used to enhance the performance of MSA algorithm. Simulation experiments were carried on 10 benchmark functions, and the results show that MMSA algorithm has good performance in terms of solution accuracy