A more accurate second-order polynomial metamodel using a pseudo-random number assignment strategy

Abstract

The response surface metamodel is a useful sequential methodology for approximating the relationship between the input variables and the output response in computer simulation. Several strategies have been proposed to increase the accuracy of the estimation of the metamodel. In the current paper, we introduce an effective pseudo-random number (PRN) assignment strategy with Box-Behnken design to construct a more accurate second-order polynomial metamodel to estimate the network reliability of a complex system. The results obtained from the simulation approach show that the reduction in maximum absolute relative error between the response surface approximation and the actual reliability function is 35.63% after the PRN assignment strategy is applied.

Keywords:

• network reliability
• response surface metamodel
• simulation
• pseudo-random number assignment

Acknowledgements

This study is partially supported by the National Science Council of Republic of China under Grant No. NSC098-2811-E-007-070. The author thanks two anonymous referees for helpful comments.