Optimal Selection of Sequential Tests for Reliability

Reliability test requirements exist whereby truncated sequential test methods offer management the opportunity to minimize the s-expected cost of a test by optimal selection of the blend of resources that will be consumed by the test. The models presented here are useful in the real world design of reliability tests because of the ease of arriving at cost estimates and the existing availability of statistical distributional information on truncated sequential tests. The test parameters of sample size, number of spares, and duration of the test are amenable to optimization for reduction in the cost of reliability acceptance testing. When the various costs associated with the test are known, and the pdf and Sf of failures and/or time to test termination are defined, then minimum s-expected test costs are achievable. Constraints and sensitivity analysis can be applied to the models to assist the managerial decision process.