Software Failure Data Research Articles

Hybrid Software Reliability Model for Big Fault Data and Selection of Best Optimizer Using an Estimation Accuracy Function

Software reliability analysis has come to the forefront of academia as software applications have grown in size and complexity. Traditionally, methods have focused on minimizing coding errors to guarantee analytic tractability. This causes the estimations to be overly optimistic when using these models. However, it is important to take into account non-software factors, such as human error and hardware failure, in addition to software faults to get reliable estimations. In this research, we examine how big data systems' peculiarities and the need for specialized hardware led to the creation of a hybrid model. We used statistical and soft computing approaches to determine values for the model's parameters, and we explored five criteria values in an effort to identify the most useful method of parameter evaluation for big data systems. For this purpose, we conduct a case study analysis of software failure data from four actual projects. In order to do a comparison, we used the precision of the estimation function for the results. Particle swarm optimization was shown to be the most effective optimization method for the hybrid model constructed with the use of large-scale fault data.

Using the Methods of Statistical Data Analysis to Improve the Trustworthiness of Software Reliability Modeling

In the past, several <i>software reliability growth models</i> (SRGMs) have been proposed to predict various aspects of software reliability, and non-homogeneous Poisson process (NHPP)-based SRGMs have drawn much attention. However, it should be noted that, in practice, blindly applying SRGMs might not lead to meaningful results when the characteristics of the data (i.e., data type) do not fit the characteristics of selected SRGMs. Moreover, there currently exist no guidelines for project managers or developers to follow in selecting the most appropriate SRGMs for a particular failure dataset. On the other hand, waiting to collect a substantial amount of failure data before fitting SRGMs might be unfeasible in many cases. In this paper, we propose useing statistical data analysis methods to investigate the characteristics of software failure data and to improve the accuracy of software reliability growth modeling. Here, SRGMs will be broadly partitioned into two categories: &#x201C;NHPP models&#x201D; and &#x201C;non-NHPP models.&#x201D; Similarly, software failure data can be classified into two categories: &#x201C;NHPP-type data&#x201D; and &#x201C;non-NHPP-type data.&#x201D; We propose a two-phase method to verify the hypothesis that failure data follow an NHPP. Through the proposed two-phase and Laplace trend tests, we will be able to infer whether a dataset is the NHPP type and whether it is properly used by NHPP-based SRGMs. In this case, the final choice for the most suitable NHPP model for the software failure process would be made, and the software reliability prediction result can be deemed trustworthy and helpful after correctly judging whether the dataset can be classified or not as being NHPP based. Experiments based on 25 real software failure data were presented and analyzed. Using our proposed method, project managers and/or developers will be able to analyze and use collected failure data to help improve both planning accuracy and software quality.

SRGM Dependent Learning based Testing Effort for Refinement

Abstract: During the development of software code there is a pressing necessity to remove the faults or bugs and improve software reliability. To get the accurate result, in every phase of software development cycle assessments needs to be happen, so that in each phase early bugs detection takes place that leads to maintain accuracy at each level. The academic institutions and industries are enhancing the development techniques in software engineering and their by performing regular testing for finding faults in programmers of software during the development. New programs are composed by altered the original code by comprised more of a bias near statements that arise in pessimistic execution paths. Fault localization information technique is used in proposed method to indicate the position of fault. In experimental as well as regression based equations represent the soft computing techniques results is better compare to the other techniques. Evaluation of soft-computing techniques represented that accuracy of the ANN model is superior to the other models. Data bases for performing the training and testing stages were collected, these soft computing techniques had low computational errors than the empirical equations. Finally says that soft computing models are better compare to the regression models. Hence, finding faults and correcting a serious software problem would be better instead of recalling thousands of products, especially in automotive sector. SRGM success mainly reliable by gathering the accurate failure information. The functions of the software reliability growth model were predicted in terms of such information gathered only. SRGM techniques in the literature and it gives a reasonable capability of value for actual software failure data. Therefore, this model, in future, can be applied to operate a wide range of software and its applications. Keywords: SRGM, FDP, FCP

Software reliability prediction and release time management with coverage

PurposeSoftware testing is needed to produce extremely reliable software products. A crucial decision problem that the software developer encounters is to ascertain when to terminate the testing process and when to release the software system in the market. With the growing need to deliver quality software, the critical assessment of reliability, cost of testing and release time strategy is requisite for project managers. This study seeks to examine the reliability of the software system by proposing a generalized testing coverage-based software reliability growth model (SRGM) that incorporates the effect of testing efforts and change point. Moreover, the strategic software time-to-market policy based on costreliability criteria is suggested.Design/methodology/approachThe fault detection process is modeled as a composite function of testing coverage, testing efforts and the continuation time of the testing process. Also, to assimilate factual scenarios, the current research exhibits the influence of software users refer as reporters in the fault detection process. Thus, this study models the reliability growth phenomenon by integrating the number of reporters and the number of instructions executed in the field environment. Besides, it is presumed that the managers release the software early to capture maximum market share and continue the testing process for an added period in the user environment. The multiattribute utility theory (MAUT) is applied to solve the optimization model with release time and testing termination time as two decision variables.FindingsThe practical applicability and performance of the proposed methodology are demonstrated through real-life software failure data. The findings of the empirical analysis have shown the superiority of the present study as compared to conventional approaches.Originality/valueThis study is the first attempt to assimilate testing coverage phenomenon in joint optimization of software time to market and testing duration.

Measuring software reliability under the influence of an infected patch

Patching service provides software firms an option to deal with the leftover bugs and is thereby helping them to keep a track of their product. More and more software firms are making use of this concept of prolonged testing. But this framework of releasing unprepared software in market involves a huge risk. The hastiness of vendors in releasing software patch at times can be dangerous as there are chances that firms release an infected patch. The infected patch (es) might lead to a hike in bug occurrence and error count and might make the software more vulnerable. The current work presents an understanding of such situation through mathematical modeling framework; wherein, the distinct behavior of testers (during in-house testing and field testing) and users is described. The proposed model has been validated on two software failure data sets of Tandem Computers and Brazilian Electronic Switching System, TROPICO R-1500.

A generalized Gompertz growth model with applications and related birth-death processes

In this paper, we propose a flexible growth model that constitutes a suitable generalization of the well-known Gompertz model. We perform an analysis of various features of interest, including a sensitivity analysis of the initial value and the three parameters of the model. We show that the considered model provides a good fit to some real datasets concerning the growth of the number of individuals infected during the COVID-19 outbreak, and software failure data. The goodness of fit is established on the ground of the ISRP metric and the d_2-distance. We also analyze two time-inhomogeneous stochastic processes, namely a birth-death process and a birth process, whose means are equal to the proposed growth curve. In the first case we obtain the probability of ultimate extinction, being 0 an absorbing endpoint. We also deal with a threshold crossing problem both for the proposed growth curve and the corresponding birth process. A simulation procedure for the latter process is also exploited.

Failure prediction by regularized fuzzy learning with intelligent parameters selection

Software failure time prediction, which is critical in software dependable evaluation, is among one of the most critical issues in software quality assurance research. Conventional software reliability predictions often use single models to illustrate the overall characteristic of the process of software failure. However, those models often fail to produce good prediction performance owing to the constant changes in software failure patterns during different time periods. Fuzzy systems like Takagi–Sugeno–Kang are multimodel approaches that combine simple submodels to represent the overall characteristic of dynamic nonlinear complex systems. This paper presents a novel Fuzzy system based software reliability prediction model, called regularized extreme learning adaptive neuro-fuzzy inference system, and employs the quantum-inspired binary gravitational search algorithm to determine its suitable regularization parameters. The new model can effectively reduce randomness, computational complexity as well as avoid local optimization when searching for optimal parameters. An experimentation using four real-world software failure data sets was carried out to compare the performance of the proposed model with that of certain representative software reliability prediction models. The results showed that the proposed model with the optimized parameters exhibits superior performance.

Research on the software reliability quantitative evaluation of nuclear power plant digital control system based on non-homogeneous poisson process model

With the application of digital technology, the software used in various systems has become increasingly important, as well as complex. This increase in complexity and scale has led to an increase in the probability of software failure. Quantitative reliability evaluation of the digital instrumentation and control system software used in nuclear power plants represent a significant challenge. This study uses a non-homogeneous Poisson process model for quantitative evaluation of the software reliability of a nuclear power plant safety digital control system (DCS), which is the research object. To analyse the research object, a software test environment is established, and the research object is tested to obtain the software failure data. The software reliability growth model, which is widely used in other industries, is combined with data on nuclear power plant DCS software defects that obey a non-homogeneous Poisson process, to construct a non-homogeneous Poisson process model. The formulas for software reliability indicators such as total failures, remaining failures, failure rate (failures per week), reliability, and time to next failure(s) are derived. The software failure data obtained from the test are used to determine the variables of the model based on maximum likelihood estimation; the estimates are then used to quantitatively evaluate software reliability-related indicators. Finally, the minimum correlation error method is used to analyse the error by comparing the actual and model prediction results. The model can evaluate the current reliability level of software and predict the impact of subsequent reliability test results. The findings from this study can provide data support for evaluating the reliability of nuclear power plant DCS software. Therefore, this study can be significant for further research on the reliability of nuclear power plant DCS software.

Software reliability prediction and management: A multiple change‐point model approach

AbstractIt is commonly recognized that software development is highly unpredictable and software quality may not be easily enhanced after software product is finished. During the software development life cycle (SDLC), project managers have to solve many technical and management issues, such as high failure rate, cost over‐run, low quality, and late delivery. Consequently, in order to produce robust and reliable software product(s) on time and within budget, project managers and developers have to appropriately allocate limited time, manpower, development, and testing effort. In the past, the distribution of testing effort or manpower can typically be described by the Weibull or Rayleigh model. Practically, it should be noticed that development environments or methods could be changed due to some reasons. Thus, when we plan to perform software reliability modeling and prediction, these changes or variations occurring in the development process have to be taken into consideration. In this paper, we will study how to use the Parr‐curve model with multiple change‐points to depict the consumption of testing effort and how to perform further software reliability analysis. Some mathematical properties of proposed model will be given and discussed. The applicability and performance of our proposed model will be demonstrated and assessed through real software failure data. Experimental results are analyzed and compared with other existing models to show that our proposed model gives better predictions. Finally, an optimal software release policy based on cost‐reliability criteria is proposed and studied. The main purpose is aimed at minimizing the total cost of software development when a reliability objective is given.

Dynamic testing resource allocation modeling for multi-release software using optimal control theory and genetic algorithm

PurposeThe use of software is overpowering our modern society. Advancement in technology is directly proportional to an increase in user demand which further leads to an increase in the burden on software firms to develop high-quality and reliable software. To meet the demands, software firms need to upgrade existing versions. The upgrade process of software may lead to additional faults in successive versions of the software. The faults that remain undetected in the previous version are passed on to the new release. As this process is complicated and time-consuming, it is important for firms to allocate resources optimally during the testing phase of software development life cycle (SDLC). Resource allocation task becomes more challenging when the testing is carried out in a dynamic nature.Design/methodology/approachThe model presented in this paper explains the methodology to estimate the testing efforts in a dynamic environment with the assumption that debugging cost corresponding to each release follows learning curve phenomenon. We have used optimal control theoretic approach to find the optimal policies and genetic algorithm to estimate the testing effort. Further, numerical illustration has been given to validate the applicability of the proposed model using a real-life software failure data set.FindingsThe paper yields several substantive insights for software managers. The study shows that estimated testing efforts as well as the faults detected for both the releases are closer to the real data set.Originality /valueWe have proposed a dynamic resource allocation model for multirelease of software with the objective to minimize the total testing cost using the flexible software reliability growth model (SRGM).

Effort based release time of software for detection and correction processes using MAUT

The amalgamation of technologies in the information technology sector resulted in unprecedented growth in software reliability modelling in the past four decades. Over the period much attention has been given on developing a fault detection model and then considering fault correction as debugging processes as a complex process in which a detected fault may not immediately be corrected because of manipulation delays and some uncertainties. To overcome these issues academician and industrialists both are working hand in hand to find out the optimal software launch time by minimizing overall testing cost. Through this work, we have developed a unified framework for developing testing effort-based software reliability growth models to optimize the launch time of the software by minimizing the total cost incurred during the software lifecycle and considering fault detection and correction separately. Further, we have used multi-attribute utility theory by considering four attributes viz. cost, the ratio of detected and corrected faults, the ratio of the corrected and total number of faults, total effort utilized to establish the optimum software launch time. The recommended model is authenticated on a real-life software failure data set.

Parameter Estimation of Software Reliability Model and Prediction Based on Hybrid Wolf Pack Algorithm and Particle Swarm Optimization

Software reliability is estimated and predicted based on software reliability model and software failure data. As a new optimization method, swarm intelligence algorithm has been widely used in solving the parameter optimization of the model. WPA (Wolf Pack Algorithm) and PSO (Particle Swarm Optimization) are two typical swarm intelligence algorithms. WPA has a strong global optimization ability, fast convergence speed and various optimization strategies, but the algorithm is relatively complex. PSO algorithm has a simple structure and fast convergence speed, but it is easy to fall into premature, which leads to low accuracy of solution. Considering the advantages and disadvantages of the two algorithms, a hybrid method of WPA and PSO is proposed, and a fitness function is constructed on maximum likelihood estimation, then the parameters of software reliability model are estimated and predicted based on the hybrid algorithm (WPA-PSO). Five sets of data from industry are used to estimate the parameters of GO model and make predictions. The simulation results show that the hybrid algorithm has higher accuracy of parameter estimation, better optimization performance, better accuracy of prediction and algorithm stability than single algorithm, and show obvious advantages than the single algorithm in the case of limited data.

A New Criterion for Model Selection

Selecting the best model from a set of candidates for a given set of data is obviously not an easy task. In this paper, we propose a new criterion that takes into account a larger penalty when adding too many coefficients (or estimated parameters) in the model from too small a sample in the presence of too much noise, in addition to minimizing the sum of squares error. We discuss several real applications that illustrate the proposed criterion and compare its results to some existing criteria based on a simulated data set and some real datasets including advertising budget data, newly collected heart blood pressure health data sets and software failure data.

Forecasting of software reliability using neighborhood fuzzy particle swarm optimization based novel neural network

This paper proposes an artificial neural network &#x0028 ANN &#x0029 based software reliability model trained by novel particle swarm optimization &#x0028 PSO &#x0029 algorithm for enhanced forecasting of the reliability of software. The proposed ANN is developed considering the fault generation phenomenon during software testing with the fault complexity of different levels. We demonstrate the proposed model considering three types of faults residing in the software. We propose a neighborhood based fuzzy PSO algorithm for competent learning of the proposed ANN using software failure data. Fitting and prediction performances of the neighborhood fuzzy PSO based proposed neural network model are compared with the standard PSO based proposed neural network model and existing ANN based software reliability models in the literature through three real software failure data sets. We also compare the performance of the proposed PSO algorithm with the standard PSO algorithm through learning of the proposed ANN. Statistical analysis shows that the neighborhood fuzzy PSO based proposed neural network model has comparatively better fitting and predictive ability than the standard PSO based proposed neural network model and other ANN based software reliability models. Faster release of software is achievable by applying the proposed PSO based neural network model during the testing period.

An improved software reliability prediction model by using high precision error iterative analysis method

SummarySoftware reliability deals with the probability that software will not cause the failure of a system in a specified time interval. Software reliability growth models (SRGMs) are used to predict future behaviour from known characteristics of software, like historical failures. With the increasing demand to deliver quality software, more accurate SRGMs are required to estimate the software release time and cost of the testing effort. Software failure predictions at early phases also provide an opportunity for investing in proper quality assurance and upfront resource planning. Up till now, many parametric software reliability growth models (PSRGMs) have been proposed. However, several limitations of them mean that their predictive capacities differ from one dataset to others. In this paper, to enhance the prediction accuracy of existing PSRGMs, a high precision error iterative analysis method (HPEIAM) has been proposed based on the residual errors. In HPEIAM, residual errors from the estimated results of SRGMs are considered as another source of data that can combine the residual error modification with artificial neural network sign estimator. The repeated computation of residual errors by SRGMs improves and corrects the prediction accuracy up to the expected level. The performance of HPEIAM is tested with several PSRGMs using two sets of real software failure data based on three performance criteria. Moreover, we have compared the estimated failures predicted by HPEIAM with genetic algorithm (GA)‐based prediction improvement. The results demonstrate that HPEIAM gives an improvement in goodness‐of‐fit and predictive performance for every PSRGM in initial few iterations.

A neuro-particle swarm optimization logistic model fitting algorithm for software reliability analysis

This article develops a particle swarm optimization algorithm based on a feed-forward neural network architecture to fit software reliability growth models. We employ adaptive inertia weight within the proposed particle swarm optimization in consideration of learning algorithm. The dynamic adaptive nature of proposed prior best particle swarm optimization prevents the algorithm from becoming trapped in local optima. These neuro-prior best particle swarm optimization algorithms were applied to a popular flexible logistic growth curve as the [Formula: see text] model based on the weights derived by the artificial neural network learning algorithm. We propose the prior best particle swarm optimization algorithm to train the network for application to three different software failure data sets. The new search strategy improves the rate of convergence because it retains information on the prior particle, thereby enabling better predictions. The results are verified through testing approaching of constant, modified, and linear inertia weight. We assess the fitness of each particle according to the normalized root mean squared error which updates the best particle and velocity to accelerate convergence to an optimal solution. Experimental results demonstrate that the proposed [Formula: see text] model based prior best Particle Swarm Optimization based on Neural Network (pPSONN) improves predictive quality over the [Formula: see text], [Formula: see text], and existing model.

A Generalized Software Reliability Growth Model With Consideration of the Uncertainty of Operating Environments

This paper proposes a generalized model to cover imperfect debugging and the uncertainty of the operating environment and its effect on fault detection rate into software reliability evaluation based on a non-homogeneous Poisson process (NHPP). Many NHPP software reliability growth models (SRGMs) have been developed to estimate the software reliability measures over the past 40 years, but most of these models assume that the operating environment is the same as the testing environment. However, in fact, due to the unpredictability of the uncertain factors in the operating environments for the software, they may considerably influence the software's reliability in an unpredictable way. So when a software system works in a field environment, its reliability is usually different from the original reliability prediction in the testing phase of the software development process, also from all its similar applications in other fields. In this paper, a general model is used to derive models that incorporate the uncertainty of operating environments, which provides the flexibility in considering a different fault detection rate and random environmental factor and so on. Several published models are shown to be covered by this general model and a new model is also developed and examined. The numerical illustrative examples of the proposed model have been validated on two sets of real software failure data in terms of six criteria. The comparison results demonstrate that the new model can fit and predict significantly better than other existing models.

Using Hybrid Algorithm to Estimate and Predicate Based on Software Reliability Model

The software reliability is mainly obtained through modeling and estimating. The existing software reliability models are nonlinear, and the parameter estimation of these models is difficult. At the same time, there are many optimization methods for solving the nonlinear function problems, such as artificial bee colony algorithm (ABC) and Particle Swarm Optimization (PSO). ABC has the characteristics of fewer control parameters, strong exploration ability, and the high accuracy of the solution. The PSO algorithm has the characteristics of the relatively small amount of computation and fast search speed but it has premature convergence, especially in dealing with complex multi-peak search problems, and the problem of poor local search ability. This paper proposes the parameter estimation method of software reliability model based on hybrid PSO-ABC, constructs a new fitness function based on maximum likelihood estimation, removes the wrong solution during the algorithm execution process, and adds knowledge to improve the solution accuracy. This paper uses five classic sets of software failure data to estimate the GO model parameters and make predictions and performs a variety of comparisons of the algorithm results. The experimental results show that the new fitness function is better, the solution of parameter estimation using the hybrid PSO-ABC is more accurate, and the hybrid PSO-ABC has a great advantage in generality and especially in limited data.

Cost-Reliability-Optimal Release Time of Software with Patching Considered

Testing life cycle poses a problem of achieving a high level of software reliability while achieving an optimal release time for the software. To enhance the reliability of the software, retain the market potential for the software and reduce the testing cost, the enterprise needs to know when to release the software and when to stop testing. To achieve this, enterprises usually release their product earlier in market and then release patches subsequently. Software patching is a process through which enterprises debug, update, or enhance their software. Software patching when used as a debugging process ensures an optimal release for the product, increasing the reliability of the software while reducing the economic overhead of testing. Today, due to the diverse and distributed nature of software, its journey in the market is dynamic, making patching an inherent aspect of testing. A patch is a piece of software designed to update a computer program or its supporting data to fix or improve it. Researchers have worked in the field to minimize the testing cost, but so far, reliability has not been considered in the models for optimal time scheduling using patching. In this paper, we discuss reliability, which is a major attribute of the quality of software. Thus, to address the issues of testing cost, release time of software, and a desirable reliability level, we propose a reliability growth model implementing software patching to make the software system reliable and cost effective. The numeric illustration has been implemented using real-life software failure data set.

Assessment and Analysis of Software Reliability Using Machine Learning Techniques

Software reliability models access the reliability by fault prediction. Reliability is a real world phenomenon with many associated real time problems and to obtain solutions to problems quickly, accurately and acceptably a large no. of soft computing techniques has been developed. We attempt to address the software failure problems by modeling software failure data using the machine learning techniques such as support vector machine (SVM) regression and generalized additive models. The study of software reliability can be categorized into three parts: modeling, measurement, improvement. Programming unwavering quality demonstrating has developed to a point that important outcomes can be acquired by applying appropriate models to the issue; there is no single model all inclusive to every one of the circumstances. We propose different machine learning methods for the evaluation of programming unwavering quality, for example, artificial neural networks, support vector machine calculation approached. We at that point break down the outcomes from machine getting the hang of demonstrating, and contrast them with that of some summed up direct displaying procedures that are proportional to programming dependability models.