Using stochastic models to effectively schedule hardware test efforts

This paper identifies a stochastic method via which to plan a schedule of qualification or reliability testing of electronic hardware. The technique utilizes a Markov model to anticipate scheduling delays, exhibiting as its inherent advantages the following: (1) mean time between failures (MTBF) need not be known to generate the corresponding Markov statistics; (2) model anticipates random bottlenecks in scheduling that a deterministic model cannot; (3) model readily avails itself to computation of Markov statistics that estimate the time to completion of all tests (and its variance), in addition to the time to completion of each qualification or reliability test (and the respective variances of each test); and (4) a Markov model used in (3) has been translated into a synopsis of a proposed software configuration, to function as guideline via which to generate software that would execute the desired Markov matrix computations, in the event the time to test for a large group of DUTS is to be determined.