An exploration on debugging performance for software reliability growth models with learning effects and change-points
Kuei‐Chen Chiu · Journal of Industrial and Production Engineering · 2015
The proposed models consider multiple change-points in software testing/debugging process with time-varying learning effects and are able to reasonably describe the S and exponential-shaped debugging process, simultaneously. The proposed models included both linear and exponential learning functions in the software reliability growth models to predict the detected errors and removed errors, judge change-points by the lag time between the errors-detected and errors-removed, and discuss the parameters of learning effects with change-points in the testing process by actual data-sets. The results show those change-points usually occur when the lag time between the errors-detected and errors-removed has material change. This study also verifies the effectiveness of the proposed models with R2 and mean square error (MSE), and compares the results with those of other models using actual data-sets. The proposed models have a better fit and are more reasonable to describe the actual data.