求解非线性规划的小生境遗传模拟退火算法

Niche Genetic Simulated Annealing Algorithms for Solving Nonlinear Programming

提出了用小生境遗传模拟退火算法求解带复杂约束的非线性规划问题。首先分析了遗传算法＂早熟＂收敛以及局部搜索能力弱的不足,由此引入小生境以增加种群多样性,并抑制＂早熟＂收敛现象,同时引入模拟退火算法以增强局部搜索能力,改进进化后期收敛速度慢的不足,最后结合典型非线性规划算例验证了混合算法的效率、精度和可靠性。

Niche genetic simulated annealing algorithm is proposed for solving nonlinear programming problem with complex constraints. The defects of basic genetic algorithm, ＂premature＂ convergence and weakness of local search, are illustrated, and then niche method is introduced which can increase the diversity of population to restrain the phenomena of ＂premature＂ convergence;simulated annealing is introduced which can enhance the capacity of local search to overcome the slowness of convergence in the late evolution. Typical numerical experiments are employed to demonstrate the efficiency, high accuracy and reliability of the proposed algorithm.