Statistical Inference from Start-Up Demonstration Test Data

Abstract

New results on start-up demonstration tests are obtained. In particular, the average length of a demonstration test is derived and inferences based on demonstration test data for the probability p of a successful start-up are discussed. Practical recommendations regarding the planning of a demonstration test are made. An extension of this test is also introduced that accommodates correlation in consecutive start-ups. The procedures developed are examined through Monte Carlo simulations and illustrated with a detailed example.