【摘要】 针对实值优化问题 ,以常规进化规划为基础 ,从数学理论及生物进化机制角度出发 ,提出了一种自适应变异算子 ,设计出一种新颖的自适应进化规划算法。该算法中包含了两个重要部分 :一是采用倒指数形式来描述方差σ2 与适合度 f之间的关系 ;二是在寻优过程中 ,变异量δ自适应发生改变。数值实验表明 :这种新算法具有很强的自适应、通用性及逃脱局部极值的全局优化能力 ,且运算量小。更多还原

【Abstract】 Based on the combination of the general evolutionary programming and the mathematical theory, the paper developsa self-adaptive mutation operator and presents a new algorithm called self--adaptive evolutionary programming. The newalgorithm includes two important procedures. The first, the expression designed for variance dZ and the fitness f is inverseexponent. The second, the amount of mutation 6 is self-adaptive during the optimization. The new method is applied to somemathematical test functions and the numerical results demonstrate the strong self-adaptability and versatility of the newalgorithm.更多还原