A comparison of factorial and random experimental design methods for the development of regression and neural network simulation metamodels

摘要: This paper compares two forms of experimental design methods that may be used for the development regression and neural network simulation metamodels. The designs considered are full factorial random designs. shows that, example problems, metamodels using a randomised produce more accurate efficient than those produced by similar sized with either or networks. metamodelling techniques compared their ability to predict results from manufacturing systems have different levels complexity. comparison suggest outperform conventional metamodels, especially when data sets based on rather