【作者】 王波；

【导师】 宋俐；

【作者基本信息】 西安理工大学 ， 机械设计及理论， 2010， 硕士

【摘要】 多学科设计优化方法(Multidisciplinary Design Optimization, MDO)是为了解决复杂的多学科系统设计问题而提出的,它充分利用了工程系统中相互作用的协同机制进行系统级和学科级的优化设计。目前,它已经开始应用于国内外航空、航天、汽车、机械等多个领域。本论文介绍了多学科设计优化问题目前的研究现状及解决方法,阐述了协同优化方法的基本理论和思路,用Matlab语言编写了标准协同优化算法程序并用标准测试函数对其进行了测试及性能评价。针对标准协同算法存在的收敛不稳定的问题,本文用两种改进方法(动态松弛因子法、Kriging法)对其进行改进。测试函数的计算结果表明,改进后的协同优化算法不仅克服了存在的问题,计算收敛速度也有了明显地提高。3-RRS并联机器人的优化设计是一个非常复杂的多学科优化问题,通常很难用传统的优化算法进行求解。本论文用改进后的协同优化算法成功的分解并优化了该问题,得到了令人满意的计算结果。弹簧结构的优化设计表明,协同优化算法也能有效地解决多目标优化设计问题。更多还原

【Abstract】 Multidisciplinary Design Optimization, MDO for short, deals with design of complex system which involves many subjects, such as mechanical, electrical, hydraulic branch and so on. MDO makes the best of interactive cooperative mechanism optimizes the design of system on the systemic and subject level respectively. It has been used in many engineering fields home and abroad actually.The paper reviewed briefly the researching status on MDO, introduced methods to resolve the problem of MDO, expatiated keystone and algorithmic flow of collaborative optimization (CO), developed the program of standard CO using computer language Matlab and tested and evaluated the program by test function. Two improvements was be used on standard CO for its unstable convergence. The results of test function show that CO improved can ensure convergence steadily and fastly.The optimal design on 3-RRS parallel robot is much more complex MDO. It is difficult for normal arithmetic to resolve that. In the paper, the problem of 3-RRS parallel robot was be disassembled successfully and satisfactory computing results was obtained using CO improved.And the result of spring optimal design indicates that CO can be fit for multi-objective optimization also.更多还原