摘要
约束优化问题的传统求解方法是拉格朗日乘子法,函数的可导性和多峰性常常成为求解过程中的难题。遗传算法的并行搜索为这类问题的求解提供了一种新的途径。为了提高计算效率,有学者提出用两级遗传算法分别解决拉格朗日乘子λ及优化参数的求解问题。用多参量遗传算法可以同时解决两级优化的遗传算法,把分级优化的参数同时编码,就把两级优化转化为一级优化。经试验该算法虽不能使优化算法的计算时间大大降低,却可以使程序设计工作相对简化,同时使遗传算法程序更具通用性。
The traditional way to determine constrained optimize problem is Lagrange multiplier method. The demands of derivative and multiplication of peaks becomes troubles to this method. A new way -to determine such problems is supplied by the parallel search technique of genetic algorithm (GA). To improve the efficiency of the search, some scholars give out a method of two level search in which they determined the Lagrange multiplier X and parameters separately. A new GA method of multiple parameters is presented to improve the two level optimize GA. When the parameters which was optimized in two levels are coded at the same time, the two level optimization is converted to one level. Though this algorithm will not decrease operation time needed in our program, it can predigest the process of program designing and make it universal.
出处
《湖北汽车工业学院学报》
2003年第4期24-26,共3页
Journal of Hubei University Of Automotive Technology
关键词
约束
优化
多参量
遗传算法
constraint
optimization
multiple parameters
genetic algorithm
