Ver Post

Seminars with Professor Penev - Southampton Solent University

13 Março 2013, 14:44 - Helena Maria Lopes Romão Borges

Professor Kalin Penev
Fellow at Technology School Southampton Solent University

PhD, Nottingham Trent University, UK
MSc, Technical University of Sofia, Bulgaria
Reference publications include:
Penev, Kalin and Littlefair, Guy. (2005) Free search - a comparative analysis.
Information Sciences Journal, 172 (1-2), pp. 173-193
Penev, Kalin. (2009). Free Search – a model of adaptive intelligence.
In: Proceedings of the 2009  Int. Conf. Adaptive and Intelligent Systems. IEEE.
His current research interests focus on Adaptive heuristic methods for search and optimisation - Real value optimisation,
Global optimisation, High dimensions optimisation, Optimisation applied to real world tasks such as of communication tasks
and traffic management tasks; Computer systems overclocking  - cooling methods; Mobile Software and Applications.
 
Computer processors over-clocking: heating and cooling issues
March 21st, 14H30-16H:30, Room 1.44 Pavilhão Mecânica II
 
Abstract
Essential limitation for computer systems performance is cooling. Usually a part of the energy consumed by computer systems is transformed to heat. This heat if not managed could cause problems such as overheating or even thermal damages. First part of the presentation will overview methods for computer system and processor cooling such as air cooling, thermo-electrical cooling, liquid (water  and oil) cooling and liquid nitrogen cooling. Next part of the presentation will discuss issues related with processors cooling and the need of methods for acceleration of the thermal flow. Conclusion will focus on the role and potential benefits of computer systems generated heat reutilisation.

Adaptive heuristic methods applied to numerical optimization
March 22nd, 14H30-16H:30, Room 1.44 Pavilhão Mecânica II

Abstract
One of the challenges for modern computational intelligence is the design of algorithms capable of resolving search and optimisation tasks currently resistant to the available methods. This presentation will review real value optimisation methods such as Real Value Genetic Algorithm, Particle Swarm Optimisation, Differential Evolution and Free Search. Main focus will be on specific peculiarities of Free Search - an adaptive heuristic method for search and optimisation. Experimental results achieved by Free Search on various numerical tasks will be analysed.  Conclusion will summarise presented results.

DISCUSSION PANEL
Paulo Carreira, Professor Dep. Computer Science and Engineering IST – Taguspark
António Moreira, Professor Dep. Mechanical Engineering IST – Alameda
Ana Moita, Post-Doc Researcher Laboratory of Thermofluids, Combustion and Energy Systems, IST – Alameda
Miguel Panão, Post-Doc Researcher Laboratory of Thermofluids, Combustion and Energy Systems, IST - Alameda