¡@

Journal of Information Science and Engineering, Vol. 24 No.
5, pp. 1521-1535 (September 2008)

**Fuzzy System Reliability Analysis Using Triangular Fuzzy
Numbers Based on Statistical Data**
**Jing-Shing Yao, Jin-Shieh Su**^{*} and Teng-San Shih^{*}

*Department of Mathematics
*

National Taiwan University

Taipei, 106 Taiwan

^{*}Department of Applied Mathematics

Chinese Culture University

Taipei, 111 Taiwan

E-mail: suston@tpts8.seed.net.tw

In this article, we use the fuzzy concept to consider the reliability of serial system
and the reliability of parallel system. Since the population reliability *R*_{j} of the subsystem
*P*_{j}(*j* = 1, 2, ..., *n*) is unknown, if we use the point estimate *R*_{j} to estimate *R*_{j} from the
statistical data in the past, we don't know the probability of the error *R*_{j} - *R*_{j}. Moreover,
the reliability of the system may fluctuate around the point estimate Rj during a time
interval. It follows that to use the point estimate *R*_{j} to estimate the population reliability
Rj is not suitable for the real cases. Therefore, it is more desirable to use the statistical
confidence interval. Moreover, the probability of the error *R*_{j} - *R*_{j} can also be solved. In
this paper, we use the statistical confidence interval instead of the point estimate. We
transfer the statistical confidence interval into the triangular fuzzy number. Through
these triangular fuzzy numbers, we consider the fuzzy reliability system. We fuzzify the
reliability of both the serial and parallel systems. Through defuzzifying the fuzzy reliability
using the signed distance method; we get a fuzzy estimate of reliability in the two
systems.

*
***Keywords:**
fuzzy reliability, statistical data, signed distance, i-v fuzzy number, triangular
fuzzy number

Retrieve PDF document (**200809_15.pdf**)

Received October 31, 2006; revised March 6 & May 18, 2007; accepted June 27, 2007.

Communicated by Chin-Teng Lin.