Multiple Error Types Software Belief Reliability Growth Model Based on Uncertain Differential Equation
Zhe Liu, Rui Kang · 2021 IEEE 21st International Conference on Software Quality, Reliability and Security (QRS) · 2021
The high dependence on software in today's society has increased the demand of reliable software immediately. Many researchers have proposed various software reliability growth models (SRGMs) to forecast software reliability by analyzing failure data throughout the testing process. Unfortunately, since software is an intellectual artifact obeying cognitive informatics, its failures involve lots of epistemic uncertainty that can not be handled well by existing methods. Furthermore, different software errors have different implications and thus need different handling. In order to deal with these problems, this paper deduce a novel multiple error types software belief reliability growth model (MESBRGM) under the framework of uncertainty theory. Reliability evaluation is conducted by investigating several reliability indexes namely belief reliability and belief reliable time based on belief reliability theory. Parameter estimations for unknown parameters in MESBRGM are also derived. Real data analysis illustrates our proposed model in detail, and demonstrate its capability compared with several popular models.