PROGRAM

11th WSEAS International Conference on COMPUTERS

Agios Nikolaos, Crete Island, Greece, July 23-25, 2007

Thursday, July 26, 2007

PLENARY LECTURE 1

Island hopping with Optical Wireless: deployment and security

Professor Stamatios Kartalopoulos
University of Oklahoma, USA
E-mail: kartalopoulos@ou.edu

Abstract: Optical wireless communications, also known as Free Space Optical (FSO), is evolving from a simple point-to-point link to a mesh network. This evolution is due to the inherent features of FSO, ease of deployment (few hours or a day) compared with fiber optic networks (months or years), lower cost (simpler optical transceiver), and high bandwidth that potential can match fiber optic transmission (10 Gbps). However, FSO has its own impediments (fog and other airborne atmospheric particles), which require a good understanding so that a countermeasure strategy is developed to provide seamless service under all conditions.

In this talk, we present the salient characteristics of FSO technology and we present a FSO mesh network with RF back-up and enhanced security that offers triple-play services, and is particularly suitable for deployment in areas where network nodes form a closely spaced cluster, such as the Aegean islands. The presentation includes network scalability as well as information security strategies.

PLENARY LECTURE 2

Key Technologies in Next Generation Optical Networks

Professor Leonid Kazovsky

Photonic & Networking Research Laboratory (PNRL)

Stanford University, USA

kazovsky@stanford.edu

Abstract: Due to advances in photonics technologies and worldwide deployment of optical fibers, optical core transport networks have experienced an extraordinary increase of transmission capacity during the last decade. As long-haul optical networks and technologies were successfully deployed for inter- and cross-continental links, access networks and, to a lesser extent, metro networks emerge as the main challenge in terms of capacity, service and cost. New technologies need to be addressed for next generation optical networks include:

• Passive optical access networks evolving from TDM to WDM;

• Convergence of optical and wireless technologies;

• Efficient and transparent connectivity across access, metro and backbone networks;

• Burst-mode transmission underlying the aforementioned technologies.

This talk will discuss the opportunities and challenges in next generation optical networks, and present research results and future research directions in network architecture, routing algorithms and key enabling technologies.

PLENARY LECTURE 3

Modelling PEM Fuel Cells Using Neural Networks
 

Professor Jose Carlos Quadrado

DEEA - Instituto Superior de Engenharia de Lisboa

R. Conselheiro Emidio Navarro, 1, 1959-007 LISBOA

Portugal

E-mail: jcquadrado@deea.isel.ipl.pt

Abstract: New worldwide spread environmental legislation restricts the energy production systems to reduce the CO2 emissions. The Proton Exchange Membrane Fuel Cell (PEMFC) is the most promising technology in the energy domain, where low emissions and fast start-up is required. In the majority of its application, the PEMFC requires the behaviour control and supervision. To achieve these goals an efficient computational model is fundamental. The existing models require complete and accurate parameters’ identification, which in those systems is hard to obtain. It is possible to obtain efficient behaviour models by resorting to black-box like models that incorporate neural networks’ elements and therefore overcame the PEMFC modelling limitations.

PLENARY LECTURE 4

Information Aggregation as one of the Pillars of Intelligent Systems

Professor Imre J. Rudas

Institute of Intelligent Engineering Systems

John von Neumann Faculty of Informatics

Budapest Tech

Becsi ut 96/b

H-1034 Budapest, Hungary

Email: rudas@bmf.hu

Abstract: In intelligent systems where fuzzy rule base and reasoning mechanism are applied, like fuzzy control systems, decision support and expert systems, one of the key issues is the problem of aggregation of information represented by membership functions. Since the pioneering work of Prof. Lotfi Zadeh dated to 1965 a great number of fuzzy connectives, aggregation operators have been introduced. The results of continuous research in this area have proved that the generally used t-norms, t-conorms do not follow always the real phenomena and do not provide optimal performance. The requirement to develop more sophisticated intelligent systems demands to find new operator families. This paper summarizes some new approaches to information aggregation from the literature and the research results of the authors and his colleagues that have been carried out in recent years on generalization of conventional operators.

PLENARY LECTURE 5

Non-Linear Symbolic Program Analysis for Increased Parallelization

Professor Kleanthis Psarris

Department of Computer Science

The University of Texas at San Antonio

San Antonio, TX 78249

E-mail: psarris@cs.utsa.edu

Abstract: High end parallel and multi-core processors rely on compilers to perform the necessary optimizations and exploit concurrency in order to achieve higher performance. However, source code for high performance computers is extremely complex to analyze and optimize. In particular, program analysis techniques often do not take into account complex expressions during the data dependence analysis phase. Most data dependence tests are only able to analyze linear expressions, even though non-linear expressions occur very often in practice. Therefore, considerable amounts of potential parallelism remain unexploited. In this work we propose new data dependence analysis techniques to handle such complex instances of the dependence problem and increase program parallelization. Our method is based on a set of polynomial time techniques that can prove or disprove dependences in source codes with non-linear and symbolic expressions, complex loop bounds, arrays with coupled subscripts, and if-statement constraints. In addition our algorithm can produce accurate and complete direction vector information, enabling the compiler to apply further transformations. To validate our method we performed an experimental evaluation and comparison against the I-Test, the Omega test and the Range test in the Perfect and SPEC benchmarks. The experimental results indicate that our dependence analysis tool is accurate, efficient and more effective in program parallelization than the other dependence tests. The improved parallelization results into higher speedups and better program execution performance in several benchmarks.

PLENARY LECTURE 6

Novel Transform Domain Principal Component Analysis (PCA)

Techniques and Some Applications: Facial and Automatic Target

Recognition

Professor Wasfy B. Mikhael

Professor of Electrical and Computer Engineering

Director of Digital Signal Processing Laboratory

School Electrical Engineering and Computer Science

University of Central Florida

Orlando, FL 32816-2450

USA

Additional Authors:
Moataz M. Abdelwahab and Pradeep Ragothaman

Abstract: Novel transform domain formulations and implementations of PCA techniques will be given. It will be shown that this new approach results in considerable computational and storage savings while yielding very high accuracy. Useful applications such as Facial and Automatic Target Recognition are given confirming the considerable performance improvement.

PLENARY LECTURE 7

A New Look at Convexity, Duality, and Optimization
 

Professor Dimitri P. Bertsekas

McAfee Professor of Engineering

Massachusetts Institute of Technology

77 Massachusetts Ave.

Lab. for Information and Decision Systems

Rm 32-660D

Cambridge, MA 02139

Email: dimitrib@mit.edu

http://web.mit.edu/dimitrib/www/home.html

Abstract: This talk will review a recent book treatment of convex analysis and optimization. While the subject of the book is classical, the treatment of several of its important topics is new and in some cases relies on new research. The new lines of analysis include:

(a) A unified framework for minimax theory and constrained optimization duality as special cases of duality between two simple geometrical problems. Within this framework, the fundamental constraint qualifications needed for strong duality and existence of saddle points are quite apparent, and admit straightforward proofs.

(b) A unification of conditions for existence of solutions of convex optimization problems, conditions for the minimax equality to hold, and conditions for the absence of a duality gap in constrained optimization. This unification is based on conditions guaranteeing that a nested family of closed convex sets has a nonempty intersection.

(c) A unification of the major constraint qualifications that guarantee the existence of Lagrange multipliers for nonconvex constrained optimization. This unification is achieved through the notion of constraint pseudonormality, which is motivated by an enhanced form of the Fritz John necessary optimality conditions.

(d) The development of incremental subgradient methods for dual optimization, and the analysis of their advantages over classical subgradient methods.

SESSION: Databases and Data mining I

Chair: Jose Aguilar, Lawrence Deng

SESSION: Databases and Data Mining II

Chair: Agostino Poggi, Dong-Liang Lee

Friday, July 27, 2007

PLENARY LECTURE 8

A Concurrent Modular Neural System for Pattern Recognition and its

Applications in Biometrics and Satellite Imagery

Professor Victor-Emil Neagoe

Depart. of Electronics, Telecommunications and Information Technology,

Polytechnic University of Bucharest,

Splaiul Independentei 313, Bucharest, Romania

Email: victoremil@gmail.com

Abstract: Starting from the idea to consider an artificial neural module as a cell characterizing a specific class only, we propose and evaluate a new neural pattern recognition system , representing a collection of small neural modules, which use a global concurential strategy. Each module is used to correctly classify the patterns of one class only and the number of modules equals the number of classes. The model proved to have very good performances , both for the recognition score and also for the training time.

We also present the applications of the proposed system for biometrics (color and infrared facial image recognition) as well as for satellite imagery (recognition of multispectral pixels: vegetation, water, sand, buildings).

SPECIAL SESSION: Advanced Computational Techniques, Algorithms and Numerical Modelling, Simulation and Optimization I

Chair: Dana Simian

SESSION: Artificial Intelligence and Applications

Chair: Abdenour Labed, Victor-Emil Neagoe

SPECIAL SESSION: Advanced Computational Techniques, Algorithms and Numerical Modelling, Simulation and Optimization II

Chair: Dana Simian

SESSION: Computers and Networks

Chair: Weilian Su, Victor-Emil Neagoe

SPECIAL SESSION: Advanced Computational Techniques, Algorithms and Numerical Modelling, Simulation and Optimization III

Chair: Dana Simian

SESSION: E-services and Applications

Chair: Felicia Ionescu, Ioan Salomie

SPECIAL SESSION: Advanced Computational Techniques, Algorithms and Numerical Modelling, Simulation and Optimization IV

Chair: Dana Simian

SESSION: Image Processing

Chair: Hesham Faouk, Richard Frye

Saturday, July 28, 2007

SPECIAL SESSION: Advanced Computational Techniques, Algorithms and Numerical Modelling, Simulation and Optimization V

Chair: Dana Simian

SESSION: Operational Systems and Architecture

Chair: Stergios Papadimitriou, Zaigham Mahmood

SESSION: Information Systems and Analysis

Chair: Ioan Salomie, Lucio Tommaso De Paolis

SESSION: Security System and Cryptography

Chair: George Savii, Lucio Tommaso De Paolis

SESSION: Computers – software and Applications

Chair: Paulo Sousa, Petr Pivonka

SESSION: Computers - Decision Level and Applications

Chair: Shishir Shah, George Bebis

SESSION: Control, management and Applications

Chair: Jose Aguilar, Shishir Shah