Integration of Knowledge Forms in Design Optimization

非线性优化（NLP）技术通常用于帮助在设计过程中必须做出的决策。一旦设计中的关键变量被指定，优化就可以用来找到设计参数的最佳值。要使用非线性编程，设计师通常必须精通于认识到为什么特定技术不起作用，并选择合适的替代方案。本研究是利用非线性规划技术开发一个专家系统。该系统将基于从设计师-分析师和从优化程序的数值计算中获得的知识。专家系统将在几个层面上运行——使用适用于整个设计空间的规则，只适用于特殊情况，或者只适用于一个实例。这种专家系统的可用性将使设计人员从使用优化程序的乏味和沮丧中解脱出来。

Nonlinear optimization (NLP) techniques often are used to aid in the decisions that must be made as a design progresses. Once the critical variables in a design have been specified, optimization can be used to find the best values for parameters of the design. To use nonlinear programming, the designer must often become well-versed in recognizing why a particular technique is not working and in selecting an appropriate alternative. This research is to develop an expert system using nonlinear programming techniques. This system will be based on knowledge acquired from designers-analysts and from the numerical calculations of the optimization programs. The expert system will operate at several levels--using rules that apply to the entire design space, that apply only under special circumstances, or only in one instance. The availability of such an expert system would free a designer from much of the tedium and frustration of working with optimization programs.