Developing an adaptable sequential probability ratio test applicable for lifetime analysis of different continuous distributions. Issue 4 (4th July 2022)

• Record Type: Journal Article
• Authors: Rasay, Hasan; Alinezhad, Esmaeil
• Is Part Of: Quality technology & quantitative management. Volume 19:Issue 4(2022)
• Page Start: 511
• Page End: 530
• Publication Date: 2022-07-04
• Subjects: Lifetime -- lot -- batch -- hypothesis test -- reliability -- life testing -- acceptance sampling plans; Quality control -- Periodicals; Quality control -- Statistical methods -- Periodicals; Industrial management -- Periodicals; Industrial management; Management -- Research -- Methodology -- Periodicals; Qualitative research -- Periodicals; Management; Quality control; Quality control -- Statistical methods; Periodicals; 658.00721
• DOI: 10.1080/16843703.2021.2020954
• Languages: English
• ISSNs: 1684-3703

Abstract

ABSTRACT: It is commonly discussed by the quality/reliability engineers that acceptance sampling plans and life testing schemes designed using the concept of sequential probability ratio test ( SPRT ) can effectively decrease the cost and time of the experiment/inspection. Nevertheless, reaching a procedure to perform an SPRT -based life test usually involves some computations which are not simple and straightforward, even for simple forms of statistical distributions. Moreover, for each statistical distribution describing the model of the lifetime data, different specialized computations should be performed to propose a procedure for the life test. To address these shortcomings, we develop a novel life test according to the SPRT of the Bernoulli/binomial distribution, which can be simply, straightforwardly and effectively adapted for life testing of different continuous distributions. Our adaptable SPRT abbreviated to ASPRT is first designed considering the Weibull distribution and is then extended for gamma and other continuous distributions. In order to evaluate the performance, ASPRT is applied to different real-world and simulated data sets. To better prove the efficiency, it is also compared with a benchmark SPRT on different data sets. Both computational and comparative results demonstrate that ASPRT is able to effectively and efficiently conduct the life testing of different continuous distributions.