A modified sequential probability ratio test for truncated life tests using Log-Normal and Inverse Exponential Rayleigh distribution

Abstract

This research developed a Modified Sequential Probability Ratio Test (MSPRT) for Log Normal and Inverse Exponential Rayleigh Distributions based on truncated life testing. MSPRT is a mechanism designed to minimise the average sample size necessary for conducting statistical hypothesis tests at predetermined significance levels and power (1-β) across various maximum sample sizes (N), indicating that there exists a definitive termination threshold for the minimum number of items required to implement a sampling plan. The study findings compared the maximum sample size necessary for the test at a 5% significance level across various areas. The study concluded that as the duration of test termination increases, the maximum required sample size diminishes, while the values in the graphical representation rise more steeply with improved quality. Consequently, the proposed plan was deemed optimal for selecting the appropriate sample size. In comparison to the two tests, it is evident that the current ASN values at levels 10, 15, and 25 exceed the proposed ASN values of 9, 11, and 18, respectively. This indicates that at a specific ASN level, the proposed plan can be accepted and implemented to reduce inspection time and simultaneously save costs, thereby representing the optimal plan in relation to the existing sequential probability ratio test.