Reliability Growth Predication based on an Improved Grey Predicton Model

摘要

As limits of time, labors and expenses, observed data usually have the characteristic of small sample sizes in development test program. Redesigns or corrective actions can result in changes of reliability for equipments. We propose an improved GM(1,1) model to predict reliability growth in this paper. First, a newly initial condition in time response function is set in this improved GM(1,1) model. The newly initial condition is comprised of the first item and the last item of a sequence which is generated from applying the first-order accumulated generation operator to a sequence of raw data. Then the improved model can express the principle of new information priority well and improve prediction precision through fully applying new information in raw data. Secondly, we make use of the improved model to predict reliability growth in a numerical example. The comparison of predicted reliability growth curve from the improved GM(1,1) model and that from the Lloyd-Lipow model indicates that the improved GM(1,1) model is much better than the Lloyd-Lipow model for the reliability growth prediction.