Developing and testing algorithms for stopping testing, screening,run-in of large systems or programs

摘要

When large hardware-software systems are run-in or an acceptancetesting is made, a problem is when to stop the test and deliver/acceptthe system. The same problem exists when a large software program istested with simulated operations data. Based on two theses from theTechnical University of Denmark, the paper describes and evaluates 7possible algorithms. Of these algorithms, the three most promising aretested with simulated data. 27 different systems are simulated, and 50Monte Carlo simulations made on each system. The stop times generated bythe algorithm is compared with the known perfect stop time. Of the threealgorithms two is selected as good. These two algorithms are then testedon 10 sets of real data. The algorithms are tested with three differentlevels of confidence. The number of correct and wrong stop decisions arecounted. The conclusion is that the Weibull algorithm with 90%confidence level takes the right decision in every one of the 10 cases