MAXIMUM ENTROPY AND DISCRETE TIME SOFTWARE RELIABILITY GROWTH MODELS

Abstract

This paper provides a preliminary investigation into the application of the Maximum Entropy Principle (MEP), introduced by Jaynes in 1957, in modeling discrete time Software Reliability Growth Model (SRGM). On their own, each of these two topics are interesting with extensive applications, and here we will show how they can be combined to provide yet another application of the MEP among a huge array of proven successful applications. A brief discussion of MEP and SRGM will be given and a hitherto unnoticed relationship between MEP distribution and the Lagrange interpolation polynomials highlighted. We then show how MEP can be used to obtain some important distributions arising from discrete time SRGM. Finally, a simple example is given to illustrate the theory.