自适应分区段混合粒子群优化算法

该算法先利用Christos贪心算法将整个搜索区域进行自适应分区段,在每一区段内搜索出最优位置,然后将各区段的最优位置组成一新微粒群,继续搜索全局最优位置。而在每个区段中,又将模拟退火算法引入到粒子群优化(PSO)之中,通过Boltzmann机制选择每一区段中局部极值,使新算法在不同阶段兼顾对多样性和收敛速度的不同要求。与其他混合PSO算法相比,仿真实验表明,新算法具有较高的解精度,能较好地解决过早收敛问题。

Adaptive Muli- sections Hybrid Particle Swarm Optimization Algorithm

The algorithm divided first the whole searching region into muli-sections adaptively according to the greed algorithm of Christos, and searched the best position of each section. Then a new swarm was consisted of the best position of each section, and continued to search the best position of the global situation. In every section, it introduced the Simulated Annealing algorithm into Particle Swarm Optimization (PSO), and made the new algorithm consider the different request between diversity and convergence by using the mechanism of Boltzmann to select the local extremum of each section. Compared with other hybrid PSO algorithms, the simulation experiments indicated that the new algorithm had higher accuracy and could avoid effectively premature problem.