Library Subscription: Guest
Begell Digital Portal Begell Digital Library eBooks Journals References & Proceedings Research Collections
International Journal for Uncertainty Quantification

Impact factor: 1.000

ISSN Print: 2152-5080
ISSN Online: 2152-5099

Open Access

International Journal for Uncertainty Quantification

DOI: 10.1615/Int.J.UncertaintyQuantification.2016016194
pages 141-156

SOFTWARE RELIABILITY GROWTH MODEL WITH TEMPORAL CORRELATION IN A NETWORK ENVIRONMENT

Jiajun Xu
School of Computer Science and Engineering, Beihang University, Beijing 100191, China
Shuzhen Yao
School of Computer Science and Engineering, Beihang University, Beijing 100191, China
Shunkun Yang
School of Reliability and Systems Engineering, Beihang University, Beijing 100191, China
Peng Wang
School of Mathematics and Systems Science & International Research Institute for Multidisciplinary Science, Beihang University, Beijing 100191, China

ABSTRACT

Increasingly software systems are developed to provide great flexibility to customers but also introduce great uncertainty for system development. The uncertain behavior of fault-detection rate has irregular fluctuation and is described as a Markovian stochastic processes (white noise). However, in many cases the white noise idealization is insufficient, and real fluctuations are always correlated and correlated fluctuations (or colored noise) are non-Markovian stochastic processes. We develop a new model to quantify the uncertainties within non-homogeneous Poisson process (NHPP) in the noisy environment. Based on a stochastic model of the software fault detection process, the environmental uncertainties collectively are treated as a noise of arbitrary distribution and correlation structure. Based on the stochastic model, the analytical solution can be derived. To validate our model, we consider five particular scenarios with distinct environmental uncertainty. Experimental comparisons with existing methods demonstrate that the new framework shows a closer fitting to actual data and exhibits a more accurately predictive power.