Two-Dimensional NHPP Models Based on Several Testing-Time Functions and Their Applications
Enormous testing-time and testing-effort (central processing unit (CPU) time, execution time, man-hour, test coverage, executed test case, and so forth) are invested in testing-phase for enhancing software reliability. Although they are software reliability growth factors, existing software reliability growth models (SRGMs) are not introduced simultaneously. Also, the software reliability growth factors enable to substitute each other partly. Therefore, we develop new bivariate nonhomogeneous Poisson process (NHPP) models based on two types of testing-time functions in this paper. Concretely, we assume that the testing-time as a software reliability growth factor is composed of the testing-time and testing-effort factors. Then, the testing-time as the software reliability growth factor is expressed by the Cobb–Douglas and constant elasticity of substitution (CES) type testing-time functions. We conduct goodness-of-fit comparisons of existing models with proposed bivariate models. Also, we discuss estimation method of the optimal software release time and optimal amount of testing-effort as an application. Finally, we show numerical examples by using actual datasets.