Software reliability growth models based on Marshall-Olkin generalized exponential families of distributions

Software has been becoming an important and inevitable tool in our modern day to day life. It finds numerous applications such as space, telecommunications technology, military, nuclear plants, air traffic and medical monitoring control. To meet the continuing demand for high quality software, an enormous multitude of Software reliability Growth models have been proposed and adopted in recent years. A new family of distributions has been introduced by incorporating an additional parameter and applied to yield a new two parameter extension of exponential distribution [1]. Parikh et al. [2] call such a family of distributions as Marshall and Olkin Generalized Exponential (MOGE) distributions and studied its inferential problems. In this paper we propose NHPP software reliability Growth models based on MOGE and validate them through different metrics using real data sets.

I. INTRODUCTION

In the past few decades several Software reliability models have been proposed, modified and adopted in software industry. In spite of the diversity and elegance of many of these, there is still a need for models which can be more readily applied in practice [3]. Reliability models have been broadly classified into two categories-Fault Count models and Time Between Failure models. Under the fault count category of models, Goel and Okumoto [4] proposed the exponential growth model characterized by the following mean value function ; where the parameter is the initial fault content and indicates the fault detection rate. Following this celebrated model, many researchers have proposed several models with different mean value function ([6], [7], [8]). Under the time between failure category the models of Jelinski and Moranda [9] and Littlewood and Verrall [10] are the pioneering models. In this paper, we present a software reliability growth model based on Marshall-Olkin Generalized Exponential family of distributions proposed by Marshall and Olkin [1].