Software Reliability Growth Model with Continuous Error Domain - Application of a Linear Stochastic Differential Equation.

This paper considers the software error detection process in the testing phase, which is the final stage of the software development. The process is modeled by a stochastic process with a continuous changing variable. Employing a new mathematical technique based on the stochastic differential equation, a new software reliability growth model is constructed. First, the traditional basic differential equation which describes the average behavior in time of the number of errors, is extended to the Itô-type stochastic differential equation, and the probability distribution of the solution process is investigated. Then, using the derived distribution, several reliability assessment measures are derived. Finally, the unknown parameters of the model are estimated based on the actual test data, and the stochastic property of the temporal variation of the errors remaining in the software is examined numerically. Using MTBF (mean-time between failure), which is the important reliability assessment measure, the problem to determine the release time of the software is discussed. [ABSTRACT FROM AUTHOR]