Dear colleges, Hello! We are going to extend our MLE (Maximum Likelihood Estimation)- based tools of parameters calculation for GRP - early we solved this task only for “classic” repairable systems, such as Weibull (Good as New) and NHPP (Bad as Old). What is my problem. For GRP with constant value of restoration factor q (i.e. all events are only failures) it is fully OK, we have got correct values of scale, shape and q. But for more realistic situation, where events may be both Failures/Corrective Maintenance (CM) and Preventive Maintenance (PM), results are some strange. For comparison with initial parameters for test data set generation and results of estimations by means of other method (Bayesian/Sampling Method) our result are essentially different. Perhaps, MLE isn't appropriate for this task? May be, somebody of you has solved this task and we can discuss this situation?
Thanks beforehand. Regards, Sergey.
Dear Haresh, thanks! unfortunately, this article considers RGP with only one/constant restoration factor (i.e. failures). I should take into account both CM (failures) and PM. Regards, Sergey.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis