International Journal of Reliability, Quality and Safety Engineering

It is of great importance for software engineers and managers to evaluate software testing-progress in a large-scale software production process, since tremendous software development resources must be consumed to achieve high quality and reliability of a software product. By focusing on the behavior of the digested test-case data observed in the testing process, we construct a stochastic model and derive several quantitative measures for software testing-progress evaluation. Actual data observed in the testing process are analyzed by the proposed model, and we discuss the applicability of our models.

This paper presents a new stochastic model for determining the optimal release time for a computer software in testing phase, taking account of the debugging time lag. In the earlier works, most of software release models were considered, but it was assumed that an error detected can be removed instantaneously. In other words, none discussed quantitatively the effect of the software maintenance action in the optimal software release time. Main purpose of this work is to relate the optimal software release policy with the arrival-service process on the software operation phase by users. We use the Non-Homogeneous Poisson Process (NHPP) type of software reliability growth models as the software error detection phenomena and obtain the optimal software release policies minimizing the expected total software costs. As a result, the usage circumstance of a software in operation phase gives a monotone effect to the software release planning.

It is important to take into account the trade-off between hardware and software systems when total computer-system reliability/performance are evaluated and assessed. We develop an availability model for a hardware-software system. The system treated here consists of one hardware subsystem and one software subsystem and it is assumed that the system is down and restored whenever a hardware or a software failure occurs. Especially, for the software subsystem, it is supposed that (i) the restoration actions are not always performed perfectly, (ii) the restoration times for later software failures become longer and (iii) reliability growth occurs in the perfect restoration action. The hardware and the software failure-occurrence phenomena are respectively described by constant and geometrically decreasing hazard rates. The time-dependent behavior of the system, which alternately repeats the operational state that a system is operating without failures and the restoration state that a system is inoperable and restored, is described by a Markov process. Useful expressions for several quantitative measures of system performance are derived from this model. Finally, numerical examples are presented for illustration of system availability measurement and assessment.

In this paper, software reliability models based on a nonhomogeneous Poisson process (NHPP) are summarized. A new model based on NHPP is presented. All models are applied to two widely used data sets. It can be shown that for the failure data used here, the new model fits and predicts much better than the existing models. A software program is written, using Excel & Visual Basic, which can be used to facilitate the task of obtaining the estimators of model parameters.

A Bayesian procedure for estimating the reliability of a complex system of independent series and parallel subsystems is presented. This method accepts either binomial or Poisson test data (perhaps both or neither) as inputs, as well as prior information, at any and all levels in the system (such as the component, subsystem and system levels). Natural conjugate beta and gamma priors are considered. A numerical example concerning the unreliability of the low-pressure coolant injection system of a certain U.S. commercial nuclear-powered boiling water reactor is used to illustrate the procedure. A Mathematica^® notebook is available for implementing the method.

The 'total time on test' (TTT) concept has proven to be a very useful tool in many reliability applications. The graphical method based on this concept gives geometrical interpretations to solutions of stochastic maintenance problems as well as estimates of the optimal policies directly from empirical failure or repair data. In fact, some renewal type of maintenance models such as age replacement and burn-in problems have been analyzed by applying the TTT concept. In this paper, we discuss a repair limit replacement problem with a cost constraint for determining the optimal cost–limit to stop repairing a unit after it fails. Two solution methods for the problem are proposed: one is a straightforward/algebraic method and the other is a geometrical one based on the TTT concept. Also, we develop a computer software to calculate numerically the optimal repair-cost limit policies or its estimates by using the function of computer graphics, and refer to its educational effect throughout numerical examples.

The problem of designing systems which are composed of several redundant subsystems in series is considered. A method for determining the number of components in each subsystem which will minimize the total cost of components while maintaining a prescribed level of system reliability is derived. One advantage of the approach used here is that it is not necessary to employ an elaborate search algorithm to obtain the optimum solution.