Anu G. Aggarwal

Two Dimensional Multi-Release Software Reliability Modeling and Optimal Release Planning

Long-lived software systems evolve through new product releases, which involve up-gradation of previous released versions of the software in the market. But, upgrades in software lead to an increase in the fault content. Thus, for modeling the reliability growth of software with multiple releases, we must consider the failures of the upcoming upgraded release, and the failures that were not debugged in the previous release. Based on this idea, this paper proposes a mathematical modeling framework for multiple releases of software products. The proposed model takes into consideration the combined effect of schedule pressure and resource limitations using a Cobb Douglas production function in modeling the failure process using a software reliability growth model. The model developed is validated on a four release failure data set. Another major concern for the software development firms is to plan the release of the upgraded version. When different versions of the software are to be released, then the firm plans the release on the basis of testing progress of the new code, as well as the bugs reported during the operational phase of the previous version. In this paper, we formulate an optimal release planning problem which minimizes the cost of testing of the release that is to be brought into market under the constraint of removing a desired proportion of faults from the current release. The problem is illustrated using a numerical example, and is solved using a genetic algorithm.

Multi up-gradation software reliability growth model with faults of different severity

In todays environment of global competition where each company is trying to prove itself better than its competitors, software company have to continually do up-gradation or add-ons in their software to survive in the market. Each succeeding up-gradation offers some innovative performance enhancement or some new functionality etc distinguishing itself from the past release. But at the same time the amount of risk involved in up-gradation/add-ons of software with regard to introducing new faults or increasing the fault number in the software is also formidable. This model categorizes faults in two types: Type-1 and Type-2 (simple fault, hard fault namely) with respect to time which they take for isolation and removal after their observation. In this paper, we propose new model and new concept of multi release software development environment. The model developed is validated on real data sets for software which has been released in the market with new features.

OPTIMAL TESTING RESOURCE ALLOCATION FOR MODULAR SOFTWARE CONSIDERING COST, TESTING EFFORT AND RELIABILITY USING GENETIC ALGORITHM

The demand for complex and large-scale software systems is increasing rapidly. Therefore, the development of high-quality, reliable and low cost computer software has become critical issue in the enormous worldwide computer technology market. For developing these large and complex software small and independent modules are integrated which are tested independently during module testing phase of software development. In the process, testing resources such as time, testing personnel etc. are used. These resources are not infinitely large. Consequently, it is an important matter for the project manager to allocate these limited resources among the modules optimally during the testing process. Another major concern in software development is the cost. It is in fact, profit to the management if the cost of the software is less while meeting the costumer requirements. In this paper, we investigate an optimal resource allocation problem of minimizing the cost of software testing under limited amount of available resources, given a reliability constraint. To solve the optimization problem we present genetic algorithm which stands up as a powerful tool for solving search and optimization problems. The key objective of using genetic algorithm in the field of software reliability is its capability to give optimal results through learning from historical data. One numerical example has been discussed to illustrate the applicability of the approach.

A discrete SRGM for multi release software system

Due to demand of highly reliable software system and intense competition of the market the software developers are forced to come up with new features and enhancements in the software at a very frequent rate. Though it adds a new leash of life to the product, but it also leads to complexity and bugs in the software system. The phenomenon of adding features and functionalities to the existing products is called up-gradation. Most of the work done in software reliability engineering pertains to single release software system. These models do not take into consideration the effect of up-gradations on the subsequent releases. In this paper, we propose a discrete software reliability growth model for fault removal process of multi release software products. The proposed model is validated on real failure dataset for software with four releases. The results obtained are encouraging and fairly accurate.

The impact of bugs reported from operational phase on successive software releases

Software testing is a necessary part of software development life cycle (SDLC) to achieve a high reliable software system. In todays software environment of global competition where each company is trying to prove itself better than its competitors, software companies have to continually do up-gradation or add-ons in their software to survive in the market. Each succeeding up-gradation offers some innovative performance or new functionality, distinguishing itself from the past release. We consider the combined effect of bugs encountered during testing of present release and user reported bug from operational phase. The model developed in the paper takes into consideration the testing and the operational phase where fault removal phenomenon follows Kapur-Garg model and Weibull-model respectively. The model developed is validated on real datasets for software which has been released in the market with new features.

MODELING TWO-DIMENSIONAL SOFTWARE MULTI-UPGRADATION AND RELATED RELEASE PROBLEM (A MULTI-ATTRIBUTE UTILITY APPROACH)

The todays fast-paced, competitive environment in the field of Science and Technology, demands highly reliable hardware and software in order to achieve new breakthroughs in quality and productivity. In this scenario, first release of software products includes enough features and functionality to make it useful for the customers. Later, software companies have to come up with upgradation or add-ons in their software to survive in the market through a series of releases. Each succeeding upgradation offers some innovative performance or new functionality, distinguishing itself from the past releases. In one-dimensional Software Reliability Growth Models (SRGM) researcher used one factor such as Testing-Time, Testing-Effort or Coverage, etc. but within a two-dimensional SRGM environment, the process depends on two-types of reliability growth factors like Testing-time and Testing-effort. In addition, we also consider the combined effect of bugs encountered during testing of present release and user reported bugs from the operational phase. The model developed in the paper takes into consideration the testing and the operational phase where fault removal phenomenon follows, logistic and Weibull model, respectively. The paper also comprises of formulating an optimal release problem based on Multi-Attribute Utility Theory (MAUT). Lastly, the model validation is done on real dataset of software already released in the market with successive generations.

Optimising adoption of a single product in multi-segmented market using innovation diffusion model with consumer balking

Survival in the market calls for delighting the consumers with quality product. However, due to reasons such as changing consumer preference, negative word of mouth, and better competitive strategy the potential buyers perception and expectation towards a product before its actual adoption may be changed (known as balking). In this paper, we have incorporated this balking concept in our proposed innovation diffusion modelling. Another major factor incorporated in the model is the effect of repeat purchasers. Further in the paper, we have formulated an optimal promotional effort allocation problem for the multi segmented market which involves decisions regarding the distribution of promotional effort among different segments of the market so that the adoption of the product gets maximised. The nonlinear complex problem incorporates the effect of balking as well as repeat purchasing and is solved using genetic algorithm. Numerical example is presented for illustrating the formulation and its solution.

Modelling diffusion of successive generations of technology: a general framework

In todays market where each company is trying to prove itself better than its competitors, companies have to continually bring new generations of their products so as to survive in the market. Each succeeding generation offers some innovative performance or new functionality, distinguishing itself from the past generation. Based on the behavioural assumptions of diffusion theory, this paper proposes an extension of the Bass diffusion model that separates substitution from switching as well as leapfrogging. We also present a relationship between our model and Norton-Bass model. Empirical implications of the proposed model have been validated on data collected from two industries (semiconductor industry dynamic random access memory shipments of six generations, mainframe industry (USA)). The model describes the growth of these generations quite effectively.

Simultaneous allocation of testing time and resources for a modular software

Under the schedule pressure and resource limitations, developing a reliable software is of great concern for the software development firms. In order to capture the combined effect of these finite resources and limited testing time in determining the growth of testing progress a two dimensional modeling framework is required. In this paper we propose a two-dimensional software reliability growth model which takes into consideration both these factors in predicting number of faults removed from software. We have used Cobb Douglas production function to develop the model. Further, the proposed two dimensional modeling framework is applied for determining optimal allocation of testing time and resources simultaneously to a modular software system. There exists an extensive literature in software engineering on optimal allocation of either testing time or resources among modules. However, for a software development firm it would be more advantageous if they could concurrently allocate time and resources optimally to modular software product. In this paper we investigate such a two dimensional optimization problem which assigns testing time and manpower resources among the modules so that the total software development cost is minimized under the constraint of achieving pre defined proportion of faults removal from each module. In order to solve the formulated allocation problem a two dimensional genetic algorithm is proposed in the paper. Finally, a numerical example is presented to illustrate the formulation and solution of the allocation problem.

GENERAL FRAMEWORK FOR CHANGE POINT PROBLEM IN SOFTWARE RELIABILITY AND RELATED RELEASE TIME PROBLEM

Reliable software is the need of the hour especially as it is an indispensable part of our new technological world. Many SRGMs have been proposed considering the change point approach in literature. Change point is defined as the point where the fault detection rate changes, it happens due to number of reasons viz, proficiency of the testing team, nature of faults to be detected etc. In this paper we develop a generalized modeling framework incorporating change point discussing the changing nature of fault detection rate and form a related release time problem which minimizes the total cost of the software and maintains a desirable level of reliability. A numerical example is given for the release problem and the proposed models are validated on real software error data to show their goodness of fit and applicability.