Wqb0

2 (for where we take q1 = 0.5, b-i^ = 0.7, q2 = 0.6, b2 = 0.9, q3 = 0.5, b3 discrete time Weibull distribution, see, e.g., [Bra01], [BGX01]).

The empirical estimator and confidence interval at level 95% for the BMP-failure rate and the RG-failure rate of the system, for the total time of observation M = 5000, are given in Figure 2.

Figure 3 gives a comparison between failure rates estimators obtained for different sample sizes. We see that, as M increases, the estimators approach the true value. We also notice that the failure rates become constant as time

- true value of BMP-failure rate empirical estimator 95% confidence interval

— true value of RG-failure rate * empirical estimator

- 95% confidence interval

*********************************************************:

*********************************************************:

— true value of RG-failure rate * empirical estimator

- 95% confidence interval

- true value of BMP-failure rate empirical estimator 95% confidence interval

50 200 250 300 0 50 100 150 200 250 300

50 200 250 300 0 50 100 150 200 250 300

Fig. 2. Failure rates estimators and confidence interval at level 95%

increases, that is the semi-Markov system approaches a Markov one as time increases.

— true value of BMP-

failure rate

* empirical estimator

M

=1000

■-• empirical estimator

M

=2000

• empirical estimator

M

5000

0 50

— true value of RG-failure rate * empirical estimator: M=1000

- empirical estimator: M=2000 ■ empirical estimator: M=5000

— true value of RG-failure rate * empirical estimator: M=1000

- empirical estimator: M=2000 ■ empirical estimator: M=5000

150 200 250 300 0 50 100 150 200 250 300

0 50

150 200 250 300 0 50 100 150 200 250 300

Fig. 3. Failure rates estimators consistency

Figure 4 presents the failure rates variances estimators for some different values of the sample size. As in Figure 3, we can note the consistency of variances estimators.

References

[BBL04] Barbu, V., Boussemart, M., Limnios, N.: Discrete time semi-Markov model for reliability and survival analysis. To appear in: Communications in Statistics-Theory and Methods, 33 (11) (2004) [BL04a] Barbu, V., Limnios, N.: Discrete Time Semi-Markov Processes for Reliability and Survival Analysis - A Nonparametric Estimation Approach. In Nikulin, M., Balakrishnan, N., Meshbah, M., Limnios, N. (eds) Parametric and Semiparametric Models with

Fig. 4. Consistency of failure rates variances estimators

Applications to Reliability, Survival Analysis and Quality of Life. Birkhäuser, Boston, 487-502 (2004) [BL04b] Barbu, V., Limnios, N.: Nonparametric estimation for discrete time semi-Markov processes with applications in reliability. Submitted for publication (2004) [BMP63] Barlow, R.E., Marshall, A.W., Prochan, F.: Properties of probability distributions with monotone hazard rate. Annals of Mathematical Statistics, 34, 375-389 (1963) [BP75] Barlow, R.E., Prochan, F.: Statistical theory of reliability and life testing: probability models. Holt, Rinehart and Winston, New York (1975)

[Bra01] Bracquemond, C.:Modélisation stochastique du vieillissement en temps discret. PhD Thesis, Laboratoire de Modélisation et Calcul, Grenoble (2001)

[BGX01] Bracquemond, C., Gaudoin, O., Xie, M.: Towards a new definition of failure rate for discrete distributions. 10th International Symposium on Applied Stochastic Models and Data Analysis, Vol. 2, 266-270 (2001)

[Cse02] Csenki, A.: Transition analysis of semi-Markov reliability models -A tutorial review with emphasis on discrete-parameter approaches. In: Osaki, S. (ed) Stochastic Models in Reliability and Maintenance. Springer, Berlin, 219-251 (2002) [How71] Howard, R.: Dynamic Probabilistic systems. vol. II, Wiley, New York (1971)

[LO99] Limnios, N., Opri§an, G.: A unified approach for reliability and performability evaluation of semi-Markov systems. Appl. Stochastic Models Bus. Ind., 15, 353-368 (1999) [LO01] Limnios, N., Opri§an, G.: Semi-Markov Processes and Reliability.

Birkhäuser, Boston (2001) [LO03] Limnios, N., Ouhbi, B.: Empirical estimators of reliability and related functions for semi-Markov systems. In: Lindqvist, B., Dok-

sum, K.A. (eds) Mathematical and Statistical Methods in Reliability. World Scientific Publishing, 7 (2003) [MS00] Mode, C.J., Sleeman, C.K.: Stochastic processes in epidemiology.

World Scientific, New Jersey (2000) [MP68] Moore, E.H., Pyke, R.: Estimation of the transition distributions of a Markov renewal process. Ann. Inst. Statist. Math., 20, 411— 424 (1968)

[NO75] Nakagawa, T., Osaki, S.: The discrete Weibull distribution. IEEE

Trans. On Reliability, R-24, 300-301 (1975) [OL99] Ouhbi, B. Limnios, N.: Nonparametric estimation for semi-Markov processes based on its hazard rate functions. Statistical Inference for Stochastic Processes, 2 (2), 151-173 (1999) [OL98] Ouhbi, B., Limnios, N.: Non-parametric failure rate estimation of semi-Markov systems. In: Janssen, J., Limnios, N. (eds) Semi-Markov models and applications. Kluwer, Dordrecht, 207-218 (1998)

[PS64] Pyke, R., Schaufele, R.: Limit theorems for Markov renewal processes. Annals of Mathematical Statistics, 35, 1746-1764 (1964) [RG92] Roy, D., Gupta, R.: Classification of discrete lives. Microelectron.

Reliab., 32 (10), 1459-1473 (1992) [SL02] Sadek, A., Limnios, N.:Asymptotic properties for maximum likelihood estimators for reliability and failure rates of Markov chains. Communications in Statistics-Theory and Methods, 31 (10), 18371861 (2002)

Was this article helpful?

0 0

Post a comment