约束离散空间的试验设计方案研究
Research on Experimental Design Schemes of Constrained Discrete Space
目前的试验设计大体在连续空间或者无约束的离散空间下进行研究,但实际工程中不少场 景都需要在离散空间中挑选试验点来进行试验设计,通常的试验设计方法无法应用于约束的离散空间中.文章 立足于实际工程背景,基于连续空间下的均匀试验设计,采用Gale-Shapley 算法,给出了连续空间到离散空间的映射设计方案.最后,结合实例进行了Gale-Shapley算法与经验方 法所得设计的对比验证研究,在约束的离散空间下, Gale-Shapley 算法所得到的映射设计具有更好的均匀性,且计算效率满足工程实际需求.
At present, experimental design is roughly conducted in continuous space or unconstrained discrete space. However, in actual engineering, many experimental points need to be selected in discrete space for experimental design. The usual experimental design methods cannot be applied to constraints discrete space. Based on the actual engineering background, based on the uniform experimental design in continuous space, the Gale-Shapley algorithm is used to give a mapping design of continuous space to discrete space. Finally, a comparative verification study of the design obtained by the Gale-Shapley algorithm and the empirical method is carried out in combination with an example. In a constrained discrete space, the projection design obtained by the Gale-Shapley algorithm has better uniformity and meets the actual engineering calculation effectiveness.
约束离散空间 / 均匀试验设计 / Gale-Shapley算法 / 映射设计. {{custom_keyword}} /
/
〈 | 〉 |