An imperfect software debugging model considering irregular fluctuation of fault introduction rate

In general, software testing is a complicated and uncertain process. New faults can be introduced into the software during each fault removal. This process is called imperfect debugging. For simplicity, fault introduction rates are generally assumed to be constant. However, software debugging can be affected by many factors, such as subjective and objective influences, the difficulty and complexity of fault removal, the dependent relationships among faults, the changes in different phases of software testing, and the test schedules. Thus, the rate of fault introduction is not a constant, but is an irregularly fluctuating variable in software debugging. In this article, we propose a model with imperfect software debugging considering the irregular changes in fault introduction rates during software debugging. Experimental results reveal that our proposed model has good fitting capability and considerably stronger forecasting performance than that of the other models, and that the proposed model assumptions are close to the actual software debugging situation. Moreover, research on the irregular fluctuation of the fault introduction rate in software debugging has a certain reference value and important significance for software-intensive product testing, for instance, cloud computing.