基于遗传算法的退火精确罚函数非线性约束优化方法  

Annealing Accuracy Penalty Function Based Nonlinear Constrained Optimization Method with Genetic Algorithms

作　　者：吴志远[1] 邵惠鹤[1] 吴新余[2] 

机构地区：[1]上海交通大学自动化所 [2]南京邮电学院

出　　处：《控制与决策》1998年第2期136-140,共5页Control and Decision

基　　金：国家"九.五"攻关课题

摘　　要：提出一种新的基于遗传算法求解非线性约束优化的方法，通过自适应的退火罚因子和不可微精确罚函数来处理约束条件，可以使算法逐渐收敛于可行的极值点。仿真结果表明该方法有较高的求解精度。A new optimal method based on genetic algorithms can solve nonlinear constrained optimization problems effectively, which adopted adaptive annealing penalty factors and undifferentiable accuracy penalty function. It can use the adavantages of undifferential accuracy penalty function which can not be utilized by conventional optimal methods based on gradient. Simulation results show the algorithm having good solution precision.

关 键 词：遗传算法 非线性约束 最优化问题 惩罚函数法 

分 类 号：O224[理学—运筹学与控制论] O242.23[理学—数学]