Abstract
A new response surface methodology (RSM) called the squeeze response surface methodology (SRSM) is proposed to gain the approximate symbolic network reliability function (SNRF). The proposed SRSM can be used to solve not only complicated system configurations, but also help decision makers gain greater understanding for the structure of the system. The response value is the value of the Bonferroni bounds (using by-products of cellular automata (CA) Monte Carlo simulation (MCS), and min-cuts) minus the simulation value (obtained from CA-MCS). SRSM squeezes the range of response values to improve solution quality. Our results compare favorably with previously developed algorithms in the literature from the experiment of the benchmark example. © 2009 IEEE.