Abstract: The majority of current Latin hypercube design optimization employs the orthogonal property or homogeneous property measurement as the single optimization criterion. An experimental design method of multi-objective optimization based on modified ILS algorithm is presented to cover the shortage of single optimization criterion. The orthogonal property and homogeneous property of Latin hypercube design are optimized at the same time, and in comparison with several typical experimental design methods, its superiority is exemplified.

Key words: multi-objective optimization, modified ILS algorithm, Latin hypercube design

摘要： 现有的拉丁超立方设计优化大多以正交属性度量或均匀属性度量为单一优化准则，针对该单一优化准则存在的缺陷，提出一种基于改进ILS算法的多目标优化试验设计方法，对拉丁超立方设计的均匀属性和正交属性同时进行优化，并与已有的几种典型试验设计方法进行对比。实验结果证明了该方法的优良特性。

关键词: 多目标优化, 改进ILS算法, 拉丁超立方设计

CLC Number: 

• TP39