Abstract
It is widely recognized that software development is usually highly unpredictable and software quality may not be easily enhanced for the finished software product. During the software development life cycle, fault correction is typically costly and time-consuming. Thus the measurement and assessment of the various fault correction processes would be very important for the engineers and project managers. In the past, the queueing theory has been used to depict the software testing and debugging processes. Considering the needed time of fault correction, some infinite server queueing (ISQ) models are proposed to measure and analyze the software project. But some researches has also considered the limited number of debuggers and presented the finite server queueing (FSQ) models. However, these previous works focused only on the debugging time lag and the number of servers. If all detected faults can be classified in advance, the fault(s) can be assigned to specific (or responsible) developers to obtain better service. The aim of this paper is to apply the express-queueing theoretic approach for reducing the mean resolution time of detected faults and to support the management activities of the software project, such as staffing, cost, utilization, etc. The fault correction process will be modeled by the queueing theory, and the detected faults will also be classified and dispatched into either express queue or regular queue. Experimental results show that the overall performance and efficiency of software debugging process can be improved while using our proposed approach.