Vol. 31 · No. 5

With the advent of microarray technology, it is possible to measure gene expression levels of thousands of genes simultaneously. This helps us diagnose and classify some particular cancers directly using DNA microarray. High-dimensionality and small sample size of microarray datasets has made the task of classification difficult. These datasets contain a large number of redundant and irrelevant genes. For efficient classification of samples there is a need of selecting a smaller set of relevant and non-redundant genes. In this paper, we have proposed a two stage algorithm for finding a set of discriminatory genes responsible for classification of high dimensional microarray datasets. In the first stage redundancy is reduced by grouping correlated genes into clusters and selecting a representative gene from each cluster. Maximal information compression index is used to measure similarity between genes. In the second stage a wrapper based forward feature selection method is used to obtain a set of discriminatory genes for a given classifier. We have investigated three different techniques for clustering and four classifiers in our experiments. The proposed algorithm is tested on six well known publicly available datasets. Comparison with the other state-of-the-art methods show that our proposed algorithm is able to achieve better classification accuracy with less number of genes.

The design of Networks-on-Chip (NoCs) involves several key issues, including the topological mapping, that is, the mapping of the processing elements or Intellectual Properties (IPs) to the network nodes. Although several proposals have been focused on topological mapping last years, this topic is still an open issue. In this paper, we propose, in an extended manner, a topology-independent mapping technique for application-specific NoCs that can be used with regular or irregular topologies, and with any routing algorithm. This technique globally matches the communication pattern generated by the IPs with the available network bandwidth in the different parts of the network. The evaluation results show that the proposed technique can provide better performance than other mapping techniques not only in terms of average latency and network throughput, but also in terms of power consumption.

Color constant image description is a fundamental problem in many computer vision applications. In this paper, the diagonal-offset model is adopted as reflectance model to get color constant image descriptors. This model makes the descriptors much more robust, and also fits the real world images very well. By introducing 3D moment invariants, this paper contributes to give an illumination independent descriptor generation framework. In detail, 0-, 1- and even higher order color constant descriptors can be generated from such framework. These descriptors can characterize n-order derivative image information. Furthermore, the combination thereof can characterize not only original image but also n-order edge image color information. The experiments on real image databases show that all these descriptors are robust to illumination variation and affine transformation, and perform very well for object recognition under various situations.

A software system maintenance represents an important part of software system's lifetime. The most common reasons to change a software system are bug fixes and adding of a new functionality. Software maintenance itself is a difficult and complex process. Before applying a change, it is important to understand the software system's source code as well as the application domain. This paper presents our innovative approach to improve software system comprehension in order to simplify its maintenance. Instead of analyzing all the program code, our approach focuses on parts which are built using predefined well known software libraries. The knowledge of both – the libraries and the way they are used in software systems – allows us to identify certain concepts of the software system. This information is used to create metamodels of these concepts. The metamodel is created at a higher level of abstraction than the level of concept implementation.

In this paper we present a semantic-based approach for supporting data-intensive applications in distributed environments. The approach is characterized by usage of explicit definition of non-functional quality parameters regarding storage systems, semantic descriptions of the available storage infrastructre and monitoring data concering the infrastructure workload and users operation, along with an implementation of the approach in the form of a toolkit called FiVO/QStorMan. In particular, we describe semantic descriptions, which are exploited in the storage resource provisioning process. In addition, the paper describes results of the performed experimental evaluation of the toolkit, which confirm the effectiveness of the proposed approach for the storage resource provisioning.

The new model used to integrating security and Quality of Service (QoS) as one parameter in mobile ad-hoc network (MANET) is introduced and studied in this article. Security and QoS represent a highly important field of research in MANET and they are still being considered separately with no mechanisms used to establish cooperation between them. This new model provides alternative to cooperation between QoS and security via cross layer design (CLD) and modified security service vector. Performance analysis of the new designed model is introduced too. It is also considered herein how processing of the new integrating model affects the performance of the MANET networks.

Platform independent modeling of information systems and generation of their prototypes play an important role in software development process. However, not all tasks in this process have been covered yet, i.e. not all pieces of an information system can be designed using platform independent artifacts that are later transformable into the executable code. One of the examples is modeling of database check constraints, for which there is a lack of appropriate mechanisms to formally specify them on a platform independent level. In order to provide formal specification of check constraints at platform independent level, we developed a domain specific language and embedded it into a tool for platform independent design and automated prototyping of information systems, named Integrated Information Systems CASE (IIS*Case). In this paper, we present algorithms for transformation of check constraints specified at the platform independent level into the relational data model, and further transformation into the executable SQL/DDL code for several standard and commercial platforms: ANSI SQL-2003, Oracle 9i and 10g, and MS SQL Server 2000 and 2008. We have also implemented these algorithms in IIS*Case as a part of the process of generation of relational database schema.

This paper presents the THOMAS architecture, specially designed to model open multi-agent systems, and its application in the development of a multi-agent system for managing and planning surveillance routes for security personnel. THOMAS uses agents with reasoning and planning capabilities. These agents can perform a dynamic self-organization when they detect changes in the environment. THOMAS is appropriate for developing systems in highly dynamic environments similar to the one presented in this study, as demonstrated by the results obtained after having applied the system to a case study.

Many practical problems occur when we wish to manipulate the data in a way that requires information not included explicitly in this data, and where we have to deal with functions of such a nature. In a networked environment, the data may reside in components on a number of geographically distributed sites. These databases cannot be moved to other network sites due to security, size, and privacy consideration. In this paper, we present two self-decomposing algorithms for constructing a function from given discrete data, and finding the extrema of any function whose arguments are stored across a number of distributed databases.