International Journal of Software Engineering and Knowledge Engineering

Model Checking based verification techniques represent an important issue in the field of concurrent systems quality assurance. The lack of formal semantics in the existing formalisms describing multi-agents models combined with multi-agents systems complexity are sources of several problems during their development process. The Maude language, based on rewriting logic, offers a rich notation supporting formal specification and implementation of concurrent systems. In addition to its modeling capacity, the Maude environment integrates a Model Checker based on Linear Temporal Logic (LTL) for distributed systems verification. In this paper, we present a formal and generic framework (DIMA-Maude) supporting formal description and verification of DIMA multi-agents models.

In this article we show how knowledge representation techniques can be applied to software architecture. We define a representation model for software architecture concepts. The model is based on MY model (meta modeling in Y), which is a knowledge engineering methodology. It represents software architecture concepts using three branches: component, connector, and architecture. The component branch represents concepts that are related to computations, the connector branch represents concepts that are related to interactions, and the architecture branch represents concepts that are related to the structure and the topology of the system described. We think that such a representation of architecture concepts aids in improving reusability not only at the implementation level, but also at the description level. The model assigns a hierarchical library for the four software architecture conceptual levels (meta-meta architecture level, meta architecture level, architecture level, application level).

Application software development is normally the process of developing slightly different application models of the same application domain for different companies. Each application model corresponds to general features of the application domain and additionally involves supplementary requirements of a particular company. An application model for each individual company is always redesigned even though the general features can be reused. This leads to the need for new approaches to the development of a generic application model that can be reused with respect to a company's requirements. This paper presents a new development approach, namely Transformation-based Model Generation (TMG), which enhances the productivity of application development. It focuses on the reuse of a generic application model to automatically generate specific application models that satisfy different requirements of the same problem domain. The concepts of model representation and model generation are taken from Model Driven Architecture (MDA). XML Declarative Description (XDD) is the underlying theory for implementing the TMG. The knowledge for generating specific application models is domain-independent, hence it can be applied to various problem domains.

The major drivers for complex applications are cost, efficiency, development time, understandability, usability and more importantly the maintainability. Due to their black box nature, complexity of software components is more crucial for component-based systems. This paper discusses various complexity concerns for these systems and reviews a number of complexity metrics for software components and component-based systems. As interfaces are the only source of information to know about the black-box components, this paper proposes a new interface complexity metric for these components. This metric is based on the information available in the interfaces like interface methods and properties. It also discusses the methodology to assign the weight values to these methods and properties to finally evaluate the complexity of the component. This paper validates the proposed metric against standard Weyukar's properties and empirically evaluates the metric for several Java Bean components. Finally a correlation analysis between proposed metrics and several other metrics like performance, customizability and readability is done to validate the metric.

Many studies have compared the properties of agent-oriented methodologies, usually applying the qualitative approach. In contrast, the quantitative approach for comparing these methodologies has rarely been used. Thus, it is difficult to assess the extent of suitability of these methodologies to, and the understanding and usage of the methodologies by, their intended users. In this paper, we adopted a quantitative approach for the comparison of two agent-oriented modeling methods, the Multiagent System Engineering (MaSE) and the Object-Process Methodology for Multi-Agent Systems (OPM/MAS). In particular, we examined the modeling language of each method as reflected in the design results. The results of the examination highlighted the differences between the two methods. Specifically, MaSE was found to be more adequate for understanding interactions and in specifying a new behavior. On the other hand, OPM/MAS was found helpful in understanding the entire system and in specifying interactions with the environment. These results clearly indicate the improvements required for each of the methods compared.

It is well-known that software failure data often contain noise, making the reliability estimation problematic. In particular, the kind of data noise inherent in software failure data is biased. There is no upper bound for the value of a noisy data point, but there is a lower bound of zero. This may lead to over-optimistic estimation of the reliability when using maximum likelihood or least square methods based on standard software reliability models. We attempt to address this problem by modeling software failure data using machine learning techniques such as support vector machine regression and generalized additive models, which have mechanisms that are capable of dealing with data noise. We then analyse the results from machine learning modeling, and compare them to that of some generalized linear modeling techniques that are equivalent to standard software reliability models. The validity of the machine learning modeling of noisy software failure data is evaluated through this comparison.