The Hyper-Geometric Distribution software reliability growth Model (HGDM) was developed to estimate the number of remaining software faults after completing the test/debug phase. An important problem in the software development process is to determine when to stop testing and release the software to the users. In this paper, the cost optimal release policy, which minimizes the total expected software cost, is discussed. The total expected software cost here includes the penalty cost, which should be paid by the manufacturer if the software is delivered after the scheduled delivery time. The underlying software reliability growth model in our approach is the HGDM. Numerical examples are presented for illustration.

A considerable amount of testing resources is required during software module testing. In this paper, based on the HGDM (Hyper-Geometric Distribution Model) software reliability growth model, we investigate the following optimal resource allocation problems in software module testing: (1) minimization of the number of software faults still undetected in the system after testing given a total amount of testing resources, and (2) minimization of the total amount of testing resources repaired, given the number of software faults still undetected in the system after testing. Furthermore, based on the concepts of "average allocation" and "proportional allocation", two simple allocation methods are also introduced. Experimental results show that the optimal allocation method can improve the quality and reliability of the software system much more significantly than these simple allocation methods can. Therefore, the optimal allocation method is very efficient for solving the testing resource allocation problem.

Considerable testing resources are required during software module testing. This paper, based on the 'hyper-geometric distribution software reliability growth model' (HGDM) investigates two optimal resource allocation (OPT/RA) problems in software module testing: (1) minimization of the number of software faults (NSF) still undetected in the system after testing, given a fixed amount of testing resources; and (2) minimization of the total amount of testing resources required, given the NSF still undetected in the system after testing. Based on the concepts of average allocation and proportional allocation, two simple allocation methods are introduced. Experimental results show that the OPT/RA method can improve the quality and reliability of the software system much more than the simple allocation methods. Therefore, the OPT/RA method is very efficient for solving the 'testing resource allocation' problem.

The hyper-geometric distribution software-reliability growth model (HGDM) can estimate the number of initial faults in a software program. An important problem in software development is to determine when to stop testing and then release the software. This paper mainly investigates the optimal software release policies which minimize the mean total software cost and satisfy the software-reliability requirement based on the HGDM. The optimal software release times are determined and shown to be finite. A numerical example illustrates these optimal software release policies.

Debugging actions during the test/debug phase of software development are not always performed perfectly. That is, not all the software faults detected are perfectly removed without introducing new faults. This phenomenon is called imperfect debugging. The hyper-geometric distribution software reliability growth model (HGDM) was developed for estimating the number of software faults initially in a program. We propose an extended model based on the HGDM incorporating the notion of imperfect debugging.

The hyper-geometric distribution model (HGDM) of software reliability growth has been used for estimating the number of initial faults in a software program. Another important problem in the software development process is to determine when to stop testing and release the software. In this paper, we investigate the optimal release policies minimizing the total expected software cost with a scheduled software delivery time for the HGDM. The total expected software cost includes the penalty cost which should be paid by the manufacturer if the software is delivered after the scheduled delivery time. The main result is that the optimal release time can be determined and shown to be finite. Numerical examples illustrating the optimal software release problem are also presented.>

The hyper-geometric distribution software reliability growth model (HGDM) has been shown to be able to estimate the number of faults initially resident in a program at the beginning of the test-and-debug phase. A key factor of the HGDM is the "sensitivity factor", which represents the number of faults discovered and rediscovered at the application of a test instance. The learning curve incorporated in the sensitivity factor is generally assumed to be linear in the literature. However, this assumption is apparently not realistic in many applications. We propose two new sensitivity factors based on the exponential learning curve and the S-shaped learning curve, respectively. Furthermore, the growth curves of the cumulative number of discovered faults for the HGDM with the proposed learning curves are investigated. Extensive experiments have been performed based on two real test/debug data sets, and the results show that the HGDM with the proposed learning curves estimates the number of initial faults better than previous approaches.>