Kuei-Chen Chiu
View Full Paper
Software reliability, non-homogeneous oisson process (NHPP),learning effects, time-varying learning effects, change-points, time-varying potential errors
Recently, various software reliability growth models (SRGMs) have been proposed to assess software reliability. Some important issues in these models include the S-curve or exponential testing behaviour, the change-points in process performance, estimation of the parameters, variations in potential errors, and learning effects in the testing process. This paper provides an integrated SRGM with time-varying learning effects to deal with multiple situations of software testing/debugging, based on the non-homogeneous Poisson process (NHPP) to satisfy both S-shaped and exponential-shaped types simultaneously. An exponential learning function is adopted to describe the learning effects varied with time, and a sine function is also adopted to point out the change-points for the testing environment and the various potential errors. The results show better fit than those of other models with actual data sets. This study also verifies the effectiveness of the proposed model with R2, mean square error, and RRMS and LSE. criteria. The proposed model not only provides good numerical prediction performance for several different kinds of data but also explains the testing/debugging behaviour of the testing staff, the learning effects of the testing project itself, and the changes in the testing environment necessary to improve software system testing and management.
Important Links:
Go Back
