Abstract: A new goodness-of-fit test, K test, is developed forΓ distributions. Theoretically, the test power of K test method is higher than that of D test method, especially in the two tail regions of probabilistic density curves, which is important for high reliability design and analysis. Utilizing Monte-Carlo simulating techniques, the critical values of K test forΓ distributions with unknown scale parameters are presented.

Key words: Goodness-of-fit test, Monte-Carlo method, Reliability, Γ distribution