Reliability Growth Modeling
Ananda Sen · Wiley StatsRef: Statistics Reference Online · 2014
Abstract At the initial developmental phase of any production involving complex systems, a test‐analyze‐and‐fix (TAAF) process is often undertaken to promote the detection of system faults and incorporate appropriate redesigns or other corrective actions, and the modified system is retested. As this test–redesign–retest sequence contributes to an improvement in the system performance, failure data become increasingly sparse at the later stages of testing, making it more difficult to assess the current reliability. A reliability growth model provides a structure through which failure data from the current as well as previous stages of testing could be analyzed in an integrated way in order to make efficient inference on reliability, and other process parameters. Reliability growth models have been applied extensively in analyzing failure data arising from hardware as well as a piece of software. Further, for systems with single‐shot missions that undergo a developmental program, reliability growth models arise quite naturally while studying the effectiveness of design fixes. In this article, we describe some popular reliability growth models that have been developed in the context of observing either time to failure or sequence of polytomous outcomes.