Simultaneous Uniform DoEs for Model Building and Validation in the Multidisciplinary Optimization Framework
Ajith Narayanan, VASSILI V. TOROPOV, Alastair S. Wood, Felician Campean · 2006
In this paper two optimum Latin Hypercube (OLH) DOEs are constructed simultaneously, these are used in metamodeling as a model building set and a model validation set. In each case the goal is to optimize the uniformity of both sets with respect to the space-filling properties of the designs. A key concept is that the merged DOE, which comprises the union of the build and validation sets, also has good space-filling properties. A permutation Genetic Algorithm (PermGA) with a variety of genetic operator strategies has been implemented. The evaluation of fitness is based upon the Audze-Eglais potential energy function. The relative efficiency of various strategies for implementing the concept and the associated computational aspects are discussed with respect to the quality of designs obtained.