이산정보의 증가에 따른 아카이케정보척도 기반 신뢰성해석의 정확도 평가

Abstract

It is highly inappropriate to perform the reliability analysis for discrete information using earlier developed methods because most of these methods assumed that the distribution of variables is continuous. On the other hand reliability analysis using Akaike information criterion (AIC) is an appropriate method to calculate reliability for discrete information, since it uses discrete information directly which stands as an advantage in reliability analysis. As we know reliability becomes more accurate with increase in number of discrete information. Therefore in this paper, we suggest a suitable number of samples to calculate reliability. To evaluate accuracy of reliability, we compare AIC with Monte Carlo simulation (MCS) using mathematical examples.