TSG Reliability Practice Exam

The key idea is that Jelinski–Moranda models the failure rate as proportional to the number of latent faults still present. At any time, the instantaneous failure rate (hazard) is determined by how many latent faults haven’t been found yet. This is expressed as λ(t) = φ (N0 − m(t)), where N0 is the total number of latent faults at the start and φ is the defect-detection rate. As faults are discovered, m(t) increases, leaving fewer latent faults (N0 − m(t)), so the hazard declines over time. This structure places the model in the non-homogeneous Poisson process family but with a specific form for the rate that directly ties to remaining faults. The statement captures this defining relation, making it the best description. The alternative that describes an increasing failure rate conflicts with the decreasing hazard as faults are eliminated. The exponential mean-value form m(t) = a(1 − e^{−bt}) belongs to a different model (Goel–Okumoto). While Jelinski–Moranda can be viewed in the NHPP framework, the crucial distinguishing feature is the λ(t) = φ (N0 − m(t)) form with N0 latent faults and φ as the detection rate.