A Study on Software Reliability Growth Models Considering Imperfect Debugging

Purpose: This study develops a software reliability growth model that incorporates imperfect debugging and learning effects. The objective is to improve the accuracy and practical applicability of software reliability growth models by addressing computational difficulties commonly encountered in parameter estimation. In particular, this study emphasizes confidence interval estimation to enhance the reliability and interpretability of software reliability predictions.<br/>Methods: A simplified estimation approach is proposed to efficiently estimate model parameters and to construct confidence intervals. The proposed method provides both upper and lower bounds for the software reliability growth model. Numerical examples are presented to illustrate the implementation procedure and to demonstrate the practicality of the proposed approach.<br/>Results: The proposed model demonstrates superior performance across several evaluation criteria compared with existing models in the literature, while relying on more realistic and practical assumptions.<br/>Conclusion: This study presents a novel software reliability growth model that effectively captures imperfect debugging and learning effects. The results indicate that the proposed model serves as a valuable tool for assessing software quality and can be readily applied to real-world software systems.