Software Release Time Determination Research Articles

Machine Learning Approach for Software Reliability Growth Modeling with Infinite Testing Effort Function

Reliability is one of the quantifiable software quality attributes. Software Reliability Growth Models (SRGMs) are used to assess the reliability achieved at different times of testing. Traditional time-based SRGMs may not be accurate enough in all situations where test effort varies with time. To overcome this lacuna, test effort was used instead of time in SRGMs. In the past, finite test effort functions were proposed, which may not be realistic as, at infinite testing time, test effort will be infinite. Hence in this paper, we propose an infinite test effort function in conjunction with a classical Nonhomogeneous Poisson Process (NHPP) model. We use Artificial Neural Network (ANN) for training the proposed model with software failure data. Here it is possible to get a large set of weights for the same model to describe the past failure data equally well. We use machine learning approach to select the appropriate set of weights for the model which will describe both the past and the future data well. We compare the performance of the proposed model with existing model using practical software failure data sets. The proposed log-power TEF based SRGM describes all types of failure data equally well and also improves the accuracy of parameter estimation more than existing TEF and can be used for software release time determination as well.

Application of multi attribute utility theory in multiple releases of software

One of the important applications of software reliability models is the determination of software release time. Most of the existing studies on this topic use models based on non homogeneous Poisson process with a bounded mean value function. Multi up gradation based software reliability growth model is developed for successive release modelling and analysis. Based on this model, maximum fault removal for upgraded software and optimal release time of upgraded software are investigated as well. This paper proposes a new practical method for determining when to stop software testing considering failure intensity and cost as two factors simultaneously. This issue has been widely known as the optimal release problem of software product. The proposed new decision model based on multi-attribute utility analysis is tested on the real world data sets. In addition, the proposed decision model can help companies to make a rational decision on the optimal timing of the software.

A risk-reduction approach for optimal software release time determination with the delay incurred cost

Most existing research on software release time determination assumes that parameters of the software reliability model (SRM) are deterministic and the reliability estimate is accurate. In practice, however, there exists a risk that the reliability requirement cannot be guaranteed due to the parameter uncertainties in the SRM, and such risk can be as high as 50% when the mean value is used. It is necessary for the software project managers to reduce the risk to a lower level by delaying the software release, which inevitably increases the software testing costs. In order to incorporate the managers’ preferences over these two factors, a decision model based on multi-attribute utility theory (MAUT) is developed for the determination of optimal risk-reduction release time.

MULTI-OBJECTIVE OPTIMIZATION APPROACHES TO SOFTWARE RELEASE TIME DETERMINATION

Optimal release time determination is a typical application of software reliability models. In this decision process, reliability and cost are the two important dimensions that are generally considered. Most existing research formulates this decision process as single-objective optimization problems. Although these formulations can greatly reduce the complexity, they can hardly reflect the nature of the decision process, which is essentially a multi-objective optimization problem. More specifically, maximizing reliability and minimizing cost should be achieved simultaneously. Due to this consideration, in this paper, software release time determination is investigated by various multi-objective optimization approaches, including the trade-off analysis, multi-attribute utility theory, and physical programming approach. Numerical examples are provided for illustrative purpose. The applicability and limitations of different multi-objective optimization approaches are also discussed, which can help management to make a more reasonable decision in practice.

Risk-based software release policy under parameter uncertainty

The determination of the optimal release time is a significant problem in the software development process. Most existing research on this problem is based on the assumption that the model parameters are either known or can be accurately estimated. Due to the uncertainties associated with parameter estimation created by the very limited amount of software failure data that is generally available, the accuracy of the optimum release time determined by traditional approaches is questionable. When the mean value of the optimal release time is used, for example, there is only a 50 per cent chance that the reliability target is met at the time of release. In this paper, an optimal software release policy under parameter uncertainty is studied. To take parameter uncertainty into consideration, an optimal risk-based software release time determination approach is introduced. Application examples are given to illustrate this approach and simulation studies are carried out. The presented results can help management to consider multiple risk levels in order to reach a more reasonable decision.

Bayesian updating of optimal release time for software systems

In this paper, a Bayesian method dealing with software reliability growth with consideration of the learning effect is proposed to determine an optimal release time for software systems with regard to the testing cost and experts' prior judgments. Such an approach is able to devise an appropriate software-debugging scheme which has the best arrangement of available resources and personnel with a minimal software testing cost when lacking sufficient information for decision making. Past research on software reliability emphasized the estimation of the number of cumulative software errors or the software reliability with respect to a specific time period, yet it neglected the determination of software release time with consideration of the software testing cost, meaning that existing approaches are not entirely practical. Accordingly, the proposed method is concerned with the evaluation of the software testing cost incurred during the testing period based on experts' prior judgments and the software testing data collected within a given duration, and is thus characterized by its practicality as well as meaningfulness with consideration of the learning effect. Finally, a numerical example is given to verify the effectiveness of the proposed approach, and sensitivity and risk analyses are performed on this example.

A software reliability growth model addressing learning

Goel proposed generalization of the Goel–Okumoto (G–O) software reliability growth model (SRGM), in order to model the failure intensity function, i.e. the rate of occurrence of failures (ROCOF) that initially increases and then decreases (I/D), which occurs in many projects due to the learning phenomenon of the testing team and a few other causes. The ROCOF of the generalized non-homogenous poisson process (NHPP) model can be expressed in the same mathematical form as that of a two-parameter Weibull function. However, this SRGM is susceptible to wide fluctuations in time between failures and sometimes it seems unable to recognize the I/D pattern of ROCOF present in the datasets and hence does not adequately describe such data. The authors therefore propose a shifted Weibull function ROCOF instead for the generalized NHPP model. This modification to the Goel-generalized NHPP model results in an SRGM that seems to perform better consistently, as confirmed by the goodness of fit statistic and predictive validity metrics, when applied to failure datasets of 11 software projects with widely varying characteristics. A case study on software release time determination using the proposed SRGM is also given.

Software release time determination based on unbounded NHPP model

One of the important applications of software reliability models is the determination of software release time. Most of the existing studies on this topic use models based on nonhomogeneous Poisson process with a bounded mean value function. In this paper, nonhomogeneous Poisson process models with unbounded mean value function is used for the determination of the optimum software release time. Both reliability criteria and cost minimisation are discussed together with the used of combined criteria. Furthermore, we also consider the case of interval estimation for the software release time as point estimation is usually inaccurate.

A study of the sensitivity of software release time

Software release time determination is of great importance to software developers. In order to predict the amount of testing needed and to estimate the release time at which a certain reliability target is met, probabilistic models can be used to predict the reliability level achieved during the software testing. However, most of the results assume that the model parameters are known which is not true since they have to be estimated based on the information available. In many cases, they are inaccurate because of the small amount of data available. Hence, it is useful to know the sensitivity of software release time with respect to the estimated parameters so that attention can be paid to those parameters that affect the release time significantly. In this paper, the sensitivity issue is studied for a commonly used software reliability model. The study can be used, for example, if an overestimation of a parameter implies an underestimation of the release time which can be costly as more failures are experienced by the consumers, we should try not to overestimate that parameter. Also, if a parameter affects the release time more than others, it is important to have this parameter estimated as accurately as possible. The interval estimation of release time is recommended to avoid further excessive adjustment of release time.