Software quality estimation models generally exploit the software engineering measurements hypothesis that software metrics encapsulate the underlying quality of the software system. A typical model is trained using software measurements and fault data of a similar, previously developed project. Such a strategy requires complete knowledge of fault data for all of the training modules. However, various practical software engineering issues limit the availability of fault data for all modules in the training data. We present a semi-supervised learning scheme as a solution to software defect modeling when there is limited prior knowledge of software quality. The commonly used EM algorithm for estimating missing data values is used in conjunction with k-means clustering. An empirical investigation using software measurement and defect data from real world projects demonstrates the effectiveness and viability of the proposed method. It is shown that estimation accuracy of the defect prediction model after the semi-supervised learning process is generally better compared to a defect prediction model trained with a dataset consisting of (only available) program modules with known number of faults.
Received on March 15, 2007
References: 31

A weighted version of the k-out–of–n system is considered. In the disjoint products version of reliability analysis of weighted–k-out–of– n

systems, it is necessary to determine the order in which the weight of components is to be considered. The k–out–of–n:G(F) system consists of n components; each component has its own probability and positive integer weight such that the system is operational (failed) if and only if the total weight of some operational (failure) components is at least k. This paper designs a method to compute the reliability in computing time and in memory space. The proposed method expresses the system reliability in fewer reliability formulae than those already published.
Received on February 14, 2007
References: 07

Many real life systems such as manufacturing, telecommunication, computers, and urban traffic systems can be modeled as capacitated-flow networks. The reliability and unreliability are two important indices to measure the quality level for a capacitated-flow network. For a multiple commodity (different types of commodity) capacitated-flow network with unreliable nodes, the arcs and nodes all have several possible capacities and may fail. Different types of commodities that are transmitted through the same network simultaneously compete for the available capacities of arcs and nodes. In this paper, we first define the system capacity as a vector composed by the capacity of each component (arc or node). Then we propose a performance index, which is the probability that the upper bound of the system capacity equals a given pattern, subject to the budget constraint. The performance index can be applied to evaluate the quality level for such a network. A simple algorithm based on minimal cuts is thus presented to evaluate the performance index.
Received on July 10, 2007
References: 25

In this paper, we consider the problem of modeling and analyzing the application communication reliability of wireless sensor networks (WSN) with different topologies, including star, tree, mesh, and hierarchical clustering. We propose reliability measures that integrate the conventional connectivity-based network reliability with the sensing coverage measure of WSN. We apply a reduced ordered binary decision diagram (ROBDD) based progressive approach for evaluating the proposed coverage-oriented application communication reliability of WSN. Our study will provide useful insights for the WSN designers in choosing the appropriate network topology. We illustrate the basics and advantages of our approach by working through the analysis of an example WSN.
Received on June 28, 2007
References: 21

Deciding inspection frequency and choosing appropriate type of maintenance method are the two principal actions in maintenance program of all industries. It is well known that dynamic or variable inspection frequency is advantageous as against fixed inspection interval. For deciding the inspection schedule and maintenance type, the degradation behaviour of the deteriorating or mechanical system is a foremost criterion. Uncertainty of the information about degraded condition of the system is the hindrance to arrive at the optimal inspection/ maintenance decision. Fuzzy set theory is a viable solution to overcome such a problem. In this paper, the degraded states (degree of deterioration) of the system are modelled in the form of linguistic terms. Expert inspection engineers judge the condition based on linguistic variables that are pre-defined fuzzy sets. The opinions of the experts are then merged to arrive at a single resultant fuzzy set, which represent the present state of the deteriorating system. Distance measures and credibility score method are used to 'map' the resulting fuzzy set on to the pre-defined fuzzy sets. An illustrative maintenance decision-making strategy is devised that shows the next inspection date and the type of the maintenance to be carried out. The whole procedure is demonstrated with diverse examples.
Received on May 4, 2005
References: 13

In this paper, we discuss the software performance evaluation method considering the real-time property. The time-dependent behavior of the software system itself alternating between up and down states is described by the Markovian software availability model. Assuming that the software system can process the multiple tasks simultaneously and that the arrival process of the tasks follows a nonhomogeneous Poisson process, we analyze the distribution of the number of tasks whose processes can be completed within a prespecified processing time limit with the infinite server queueing model. We derive several stochastic quantities for software performance measurement considering the real-time property; these are given as the functions of time and the number of debugging activities. Finally, we illustrate several numerical examples of the quantities to investigate the relationship between the software reliability/restoration characteristics and the system performance.
Received on March 10, 2006
References: 16

Uncertainties are present in any reliability calculations due to randomness in the failure/repair phenomena and given the limitation in assessing the parameters of the failure/repair probability density functions. The simplifications and idealizations also generate uncertainties which can be classified as aleatory (arising due to randomness) and/or epistemic (due to lack of knowledge). It is very important to identify all the uncertainties and treat them effectively to make reliability studies more useful for decision making. The uncertainty propagation of system reliability assessment quantifies uncertainty in system characteristic (e.g. system unavailability) by synthesizing the uncertainties in component characteristics. This paper explores various uncertainty propagation methods used in availability assessment of complex engineering systems. Apart from probabilistic (analytical and sampling) and fuzzy arithmetic approaches, Dempster-Shafer theory based approach is attempted to deal the same problem. A case study on Main Control Power Supply System (MCPS) of typical Indian Nuclear Power Plant (NPP) is presented highlighting merits and demerits of various methods.
Received on Feb.11, 2007
References: 40

Erosion corrosion (EC), also referred to as Flow Accelerated Corrosion (FAC), is one of the major degradation mechanisms in Primary Heat Transport (PHT) piping system of Pressurized Heavy Water Reactor (PHWR) and therefore needs to be addressed in assessing reliability of piping system. In this paper, an existing model is used to find the FAC rate in the straight and elbow section of an outlet feeder. The effect of various factors affecting FAC is highlighted. Deterministic method is first used to evaluate the FAC rate and probabilistic method is applied to evaluate the reliability of the system under consideration. The point estimate of reliability of the piping component is illustrated, considering the components to be in series.
Received on December 31, 2006
References: 15

For the reliability analysis of systems with dynamic behavior of functional dependence, state space oriented approaches, in particular, Markov model based methods have been used for the entire system or at least dynamic portions of the system via a modular approach. Unfortunately, Markov models suffer from the well-known state explosion problem, leading to models that are computationally intensive and intractable. In this short communication, a combinatorial transform method is proposed for the reliability analysis of dynamic systems subject to functional dependencies. The method can obviate the use of Markov models, offering exact and computationally efficient solutions to the reliability analysis of large-scale dynamic systems. In addition, the proposed method will not be restricted to exponential time-to-failure distributions. The basics and advantages of the proposed approach are illustrated through an example.
Received on October 05, 2007
References: 03