When : Thursday, October 14, 2004 - 17:00
Speaker : Prof. Laurence Wolsey
Affiliation : CORE Universite Catholique de Louvain
Where : Aula Magna
Abstract :
Prof. Laurence Wolsey
CORE (Center for Operations Research and Econometrics)
Universite Catholique de Louvain
Louvain-la-Neuve, Belgium
Starting form the 50's to the 80's the area of discrete optimization
was marked by the development of linear programming, some elegant
mathematics: complexity theory and graph theory, and the relative
disappointment due to the inability to provide efficient computational
methods for solving some fundamental problems arising mainly in
engineering- management science .
Recently the pendulum has swung. Combining correct modelling with some
old and new mathematical results, remarkable improvements in linear
programming software and faster computers, it is now possible to tackle
and solve exactly or approximately large-scale problems that are
theoretically intractable and come from diverse fields: computational
biology, airline, bus and railway scheduling, electricity generation,
data mining, etc.
Recently the pendulum has swung. Combining correct modelling with some
old and new mathematical results, remarkable improvements in linear
programming software and faster computers, it is now possible to tackle
and solve exactly or approximately large-scale problems that are
theoretically intractable and come from diverse fields: computational
biology, airline, bus and railway scheduling, electricity generation,
data mining, etc.
Still, some of them remain computationally intractable, even for
relatively small dimension: This has prompted the study of alternative
methods to solve them.
In this presentation, we review a few highlights of the last 50 years,
take a couple of practical problems, and we point out some of the
challenges for the future.
Short biography
Laurence Wolsey is Professor of Applied Mathematics and Operations Research at CORE (Center for Operations Research and Econometrics) of the Engineering School of l'Universite Catholique de Louvain, Belgium.
He received a Ph.D. in mathematics from the Massachusetts Institute of
Technology.
His main field of research is integer programming, where he has given
fundamental contributions both in theory and computations.
He is author of fundamental textbooks in the domain:
'Integer Programming' 1998 and 'Integer Programming and Combinatorial
Optimization' 1988 (with G. L. Nemhauser), a reference for researchers
in the area.
He has worked with research groups at BASF (production planning),
France Telecom (multiplexer assignment) and DASH (commercial mixed
integer programming systems) among others.
He has received the Orchard-Hays prize in 1988 from the Mathematical
Programming Society (with T.J. Van Roy), the Lanchester Prize in 1989
from the Operations Research Society of America (with G.L. Nemhauser),
and the EURO Gold Medal in 1994.
He is editor-in-chief of the Mathematical Programming journal.
When : Thursday, May 27, 2004 - 17:30
Speaker : Prof. Fred C. Lee
Affiliation : The Bradley Department of Electrical and Computer Engineering, Virginia Tech, USA
Where : Aula Magna, DEI
Abstract :
FUTURE TRENDS TOWARD POWER ELECTRONICS SYSTEM INTEGRATION VIA MODULAR BUILDING BLOCK APPROACH
Fred C. Lee
The Center for Power Electronics System’s (CPES) research vision is to develop an integrated system approach via integrated power electronics modules (IPEMs). The impact of improvements on power electronics technologies and system integration via the IPEM approach can be compared to the impact being realized by improvements in very- large-scale integrated (VLSI) circuit technology. Applications of VLSI circuit technology have enabled rapid advances in information technology and the digital revolution, accompanied by a steady increase in standardization, modularization, functional integration, and a steady decrease in manufacturing costs and equipment. Parallel to this integrated circuit development, the IPEM approach makes possible increased levels of integration in the components that comprise a power electronics system – devices, circuits, controls, sensors and actuators – which are integrated into manufacturable subassemblies and modules that, in turn, can be customized for a particular application. A competitive advantage will begin by industries that can quickly and efficiently provide their customers with a level of standardization and flexibility that is now routine in VLSI circuit technology. Moreover, as processes are being established, industries that make use of IPEMS' standardized components, subassemblies and modules will be able to enjoy the savings associated with the economics of scale.
This talk will briefly discuss the technology advancements needed to improve the characteristics of power electronic systems. Among these technologies being developed are planar metallization device interconnects, allowing three-dimensional integration of power devices, as well as the integration of power passives to increase the power density as they dominate the physical size of the system. Furthermore, possible integration of EMI filters in order to minimize some of the circuit parasitics detrimental to its ability to attenuate high-frequency noises will be discussed. The technologies being developed will ultimately span a wide range of applications from distributed power systems to motor drives. The presentation will also discuss results obtained from the various technologies being developed within the research scope of CPES.
When : Friday, May 7, 2004 - 17:30
Speaker : Prof. Andrei Chi-Chih Yao
Affiliation : Computer Science Department, Princeton University, USA
Where : Aula Magna, DEI
Abstract :
Prof. Andrei Chi-Chih Yao
William and Edna Macaleer Professor of Engineering and Applied Science
Computer Science Department
Princeton University, USA
When : Friday, February 13, 2004 - 16:30
Speaker : Prof. Anton Zeilinger
Affiliation : University of Vienna, Austria
Where : Aula Magna, DEI
Abstract :
Prof. Anton Zeilinger Professor of Physics and Institute DirectorInstitute of Experimental PhysicsUniversity of Vienna, Austria
When : Friday, January 16, 2004 - 16:30
Speaker : Prof. Wolfgang Fichtner
Affiliation : ETH Zentrum, ETZ, Switzerland
Where : Aula Magna, DEI
Abstract :
The continuing advances in high performance computating have led to a fundamental paradigm shift in electronics research and development.
While time-consuming trial-and error experiments were standard practice up to several years ago, the wide-spread access to powerful computing facilities was limited, especially in industrial environments.
With the advent of cost-effective hardware platforms, the utilization of modeling concepts, based on first-principles physical foundations, offers unique advantages over the original experimental techniques.
It has become standard practice today to design materials, devices,
processing techniques and circuits 'on the computer' before committing to fabrication.
This trend towards computer experiments will certainly continue and expand in the next few years.
In this presentation, I shall review several of the interesting original developments and the current status of computational techniques in the design of micro/optoelectronic processes, devices and circuits.
I shall also attempt a view into the crystal ball to predict the future of this important and exciting field.
Prof. Wolfgang Fichtner
Head, Integrated Systems Laboratory
ETH Zentrum, ETZ, Switzerland
