An optimized method for software reliability model based on nonhomogeneous Poisson process

•It is the first time to propose an optimized method to improve the NHPP models.•The performance of optimized models is better than those of un-optimized models.•The optimized models have better robustness and adaptability. The present study proposes an optimized novel approach to improve the softwa...

Full description

Saved in:
Bibliographic Details
Published in:Applied mathematical modelling 2016-07, Vol.40 (13-14), p.6324-6339
Main Authors: Wang, Jinyong, Wu, Zhibo, Shu, Yanjun, Zhang, Zhan
Format: Article
Language:English
Subjects:
Computer programs
Faults
Mathematical analysis
Mathematical models
Nonhomogeneous Poisson process (NHPP)
Optimization
Optimized model approach
Software
Software reliability
Software reliability model
Software testing
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•It is the first time to propose an optimized method to improve the NHPP models.•The performance of optimized models is better than those of un-optimized models.•The optimized models have better robustness and adaptability. The present study proposes an optimized novel approach to improve the software reliability model based on the nonhomogeneous Poisson process (NHPP). The approach repeatedly implements the function with exponential distribution to fit a logarithmic difference between the estimated values and observed values from a software historical fault data set. Moreover, the logarithmic difference of the values gradually tends to be zero with more fittings. Furthermore, the trend in which their logarithmic difference essentially converges to a stable value over time contributes to building a fitting model and predicting the number of remaining faults in software testing. The optimal solutions are given in an optimized process. Experimental results show that the proposed optimized models fit the historical fault data set better, and more accurately predict the remaining number of faults than the traditional models based on NHPP in the software testing process.
ISSN:0307-904X
DOI:10.1016/j.apm.2016.01.016