|
|
|
| Study and application of orthogonal-genetic algorithm test method |
| MA Xiao-Li1,2, MA Lü-Zhong1, ZHOU Zhao-Zhong2 |
1. School of Mechanical Engineering, Jiangsu University, Zhenjiang 212013, China;
2. Department of Mechanical and Electronic Engineering, Quzhou College, Quzhou 324000, China |
|
|
|
Abstract To solve the problem of optimization design of parallel robot manipulators, an optimization method based on Orthogonal Experiment Method and Genetic Algorithm is proposed. On the basis of discussing the basic principles of the above two methods, compares in the aspects of the ways to hunt optimal solution and corresponding relationships among the parameters are conducted and then the method to construct the initial population of generic algorithm with orthogonal table is introduced. A method of “one generation” orthogonal-genetic algorithm test method that is applicable to greater design variables and unmanageable fit function is put forward. This method can be applied in structural optimization design of a new type of four DOF parallel robot manipulators and the optimal scheme of the mechanism’s geometric dimensions can be obtained by taking the global condition index of mechanism as objective functions. Illustrations have demonstrated this method effective.
|
|
Published: 28 December 2005
|
正交-遗传试验算法的分析与应用
为了解决并联机器人机构的优化设计问题，提出一种基于正交试验设计法和遗传算法的优化方法。在简要讨论正交试验设计法和遗传算法的基本原理基础上，对两种方法的寻优算法、各个参数的对应关系作了比较分析，探讨了用正交表构造遗传算法中初始种群的方法。提出一种适用于设计变量多且适应度函数难求的“一代”正交-遗传试验法的思路和方法。将该方法应用于一种新型四自由度并联机器人机构的结构优化设计，得出以机构全域条件数为目标函数的机构结构优化尺寸方案。实例证明这种优化方法行之有效。
关键词：
正交试验设计法,
遗传算法,
正交表,
初始种群,
并联机器人机构
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
| |
Shared |
|
|
|
|
| |
Discussed |
|
|
|
|
