摘要
利用遗传算法早熟的特点 ,构造出一种快速收敛的混合算法来求解优化问题 ,并分析了它的收敛性。它是使用遗传算法来生成搜索方向 ,从而保证了算法的收敛性。该算法利用遗传算法的全局搜索能力 ,并采用 Nelder- Mead单纯形法来加强算法的局部搜索能力 ,加快了算法的收敛速率。模拟实验表明 。
Premature convergence and low converging speed are the distinct weaknesses of genetic algorithms. A hybrid algorithm that can quickly converge to the optimal set is proposed and its convergence is analyzed. Some hybrid genetic algorithms use the genetic algorithms as the main body and directly act on the solution space of the problem. They are different from the hybrid algorithm, because the hybrid algorithm implements indirect search, that is, the search direction is generated by using GAs. On the one hand, the global search capability of GAs is utilized to guarantee the convergence of the hybrid algorithm. On the other hand, Nelder-Mead Simplex is used to strong the local search and fast convergence of the hybrid algorithm. Computed results and theory analysis indicate that the method is a robust and efficient algorithm with global optimization.
出处
《控制与决策》
EI
CSCD
北大核心
2002年第1期19-23,共5页
Control and Decision
基金
国家自然科学基金项目 (79770 0 6 0 )
