利用全局与局部单调性分析的优化设计技术

H.L.Li 和 P.Papalambros 于1984年提出一种在优化设计过程中引入人工智能的设计方法,构造一个产生式系统,用来求优化设计问题的全局最优解或者提高解为全局最优的可能性。本文在其基础上提出了一种利用全局与局部单调性分析的直接优化设计方法,用这一方法可以很容易地去掉更多的多余的子问题,简化问题的求解并进一步增加解为全局最优的可能性。此外,在一般情况下不需要求助于传统的优化技术就能直接求得问题的最优解。本文还给出了两个实际设计算例,显示出本方法的有效性。

An Optimum Design Technique Using Global and Local Monotonicity Analysis

H.L.Li and P.Papalambros have introduced a method employing ideas from applied artificial intelligence in design optimization.The method is used to identify the global optimum or increase the likelihood of identifying one by reducing the original problem to a set of smaller subproblems one of which will contain the global solution.In this paper an improved design procedure has been developed.By seeking(reducing)the new varying interval of each variable and using monotonicity analysis,the problem can be solved directly.The method presented here uses more global and local monotonicity informations and does not need the help of conventional numerical optimization techniques in general.It can also delete more redundant subproblems and thus will greatly simplify the calculating and further increase the likelihood of identifying global optimum. Two design examples,optimum design of a speed reducer and a feed gearbox of X2010 planer-type milling machine,aregiven which demonstrate the strategy and efficiency of the procedure.