| Jie Sun | |
| Jie Sun | |
| Xinmin Yang | |
| Yaroslav D. Sergeyev | |
| Alex Rubinov | |
| Gerhard-Wilhelm Weber | |
| Xinmin Yang | |
| Jiming Peng | |
| Xiaoqi Yang | |
| Defeng Sun | |
| Victor Korotkich and Panos Pardalos |
|
| Xiaoqi Yang | |
| Tamaki Tanaka | |
| Lucien Polak | |
| Liqun Qi | |
| David Q. Mayne | |
| Yorai Wardi | |
| Andre Tits | |
| Boris Polyak | |
| Xinmin Yang |
Beginning April, 1999 optimization researchers in National University of Singapore (NUS), Republic of Singapore, and Massachusetts Institute of Technology (MIT), USA, have a new platform to work together in both teaching and research, which is funded by the Singapore Government and is called The Singapore-MIT Alliance (SMA).
SMA is sort of a new graduate school in NUS , which uses INTERNET to build "the world classroom". Currently about a hundred of faculty members from both institutions work part-time in five majors involving biological science, material science, computer science, manufacturing, and high performance computing. Students from both sides of the earth take class in the same time through INTERNET and the faculties work on research projects by frequent short-term visiting. For those who can read Chinese, please refer to my paper on World Economic Review 2 (2001) for more details of the operation of SMA.
Currently, math programers are mainly involved in the High Performance Computing Program of SMA although some excellent operations researchers such as the Editor of Operation Research, Professor Wein of MIT, are participating in the manufacturing program.
Among the names familiar to us I could list the following in SMA-High performance Computing Program: Professors Tom Magnanti, Rob Freund, Dimitris Bertsimas, Jim Orlin, Georgia Perakis, Andrea Schultz, Toh Kim Chuan, Teo Chun Piaw, Huang Hui Chun, Drs Chen Xiongda, Li Rongheng, Zhou Guanglu and myself are all fellows, associates, or research fellows of this program. SMA is expected to help much in optimization research, in particular computational optimization research, in Singapore and in this region at large.
Dr Guanglu Zhou, a previous PhD student and research associate of Prof Liqun Qi, has assumed a new position as Research Fellow at the Singapore-MIT Alliance of National University of Singapore. He will closely work with Professors K-C. Toh, G. Zhao, D. Sun and J. Sun on optimization research projects.
A board member of POP, Professor J. Sun, is visiting MIT and University of Washington, USA, for Feb-June, 2001 and University of Notre Dame de la Paix, Belgium, for July-November, 2001 on his sabbatical leave from National University of Singapore (NUS). He continues taking an active part in activities of POP and NUS during his travel and can be reached via email at his NUS address: jsun@nus.edu.sg.
Global Optimization with Non-Convex Constraints:
Sequential and Parallel Algorithms
Roman G. Strongin
Nizhni Novgorod State University, Russia
Yaroslav D. Sergeyev
Institute of Systems Analysis and Information Technology of
the CNR,
c/o University of Calabria, Italy, and Nizhni Novgorod State
University, Russia
NONCONVEX OPTIMIZATION AND ITS APPLICATIONS
Volume 45
This book presents a new approach to global non-convex constrained optimization. Problem dimensionality is reduced via space-filling curves. To economize the search, constraint is accounted separately (penalties are not employed). The multicriteria case is also considered. All techniques are generalized for (non-redundant) execution on multiprocessor systems.
Audience: Researchers and students working in optimization, applied mathematics, and computer science.
Kluwer Academic Publishers, Dordrecht
Hardbound, ISBN 0-7923-6490-2
October 2000, 728 pp.
NLG 560.00 / USD 275.00 / GBP 174.00
For a complete table of contents please access Kluwer Acadmic Publisher's web site http://www.wkap.nl/. You may also order books directly on line by accessing the web.
For further details, please click here.
At July 12-13, 2001, after "EURO 2001 - The European Operational Research Conference" (9-11 7.2001), the workshop "Smooth and Nonsmooth Optimization - Theory and Applications'' of EURO working group EUROPT on continuous optimization took place in Rotterdam. This was the working group's second workshop. Since its "birth'' at the first workshop (Budapest, 2000), Prof. Dr. T. Terlaky (Hamilton, Canada; coordinator) and Prof. Dr. T. Illés (Budapest, vice coordinator) are chairing the group. This year's workshop was organized by Prof. Dr. J. Brinkhuis, Prof. Dr. H. Frenk (both from Rotterdam), Prof. Dr. T. Illés (Budapest), Prof. Dr. G. Still (Twente, Niederlande) and PD Dr. G.W. Weber (Chemnitz). The Netherlands are world-open minded. This is also reflected by the 36 person list of participants from 13 nations.
In his opening address at July 12, Prof. Dr. H. v. Dijk, director of Econometric Institute of Erasmus University, underlined the importance of optimization for economy. Then, by his invited talk, Prof. Dr. H. Th. Jongen (Aachen) introduced into the scientific programme. In terms of graphs, his topological explanations gave a "discrete guide'' in global optimization. In her contribution, Prof. Dr. C. Segastizabal (Rio) presented a new algorithm for nonsmooth optimization. Dr. J. Scheffran (Potsdam) utilized methods from vector optimization within the theory of controlled dynamical games, before Prof. Dr. M. V. Solodov (Rio) applied the theory of 2-regularity on NCP and MPEC. By her perturbation theoretical reflexions about Tikhomirov's algorithm, Dr. M. Brohé (Liège) concluded the first morning.
Next two speakers considered the case of nonsmoothness again: The invited talk of Prof. Dr. A. Kruger (Minsk) presented necessary and sufficient conditions for an enlarged optimality criterion; Prof. Dr. B. Kummer (Berlin) discussed equivalence of several conditions of Lipschitz continuity for setvalued mappings. In the next invited talk, Prof. Dr. V. Protassov (Moscow) reviewed the state of research on free-boundary problem in stochastic optimization. A macroeconomical ecological motivation underlied the time-discrete model treated by Dr. St. W. Pickl (Cologne); its optimal control led to nonsmooth functions. Analyzing marginal functions, Dipl.-Math. A. Geletu Selassi (Ilmenau) numerically approached generalized semi-infinite optimization, before Prof. Dr. T. Illés (Budapest) translated a problem of colour measurement into optimization theory and numerically solved it.
The scientific programme of July 13 began with the invited talk of Prof. Dr. P. Gritzmann (Munich). Studying some approximation problems, he took a closer look at the fruitful interface between nonlinear optimization, discrete optimization and combinatorial geometry. Dr. O. Stein (Aachen) came back to generalized semi-infinite optimization; after an analysis of such problems, he also treated them numerically. Next, we had a session on semidefinite optimization and interior point methods: While Dr. P. Huhn (Augsburg) was concerned with the dependence from the starting point and probabilistic complexity, Prof. Dr. T. Terlaky (Hamilton) introduced self-regular functions into consideration. In his invited talk, Prof. Dr. C. Roos (Delft) presented a new primal-dual method.
The series of afternoon lectures was opened by Prof. Dr. R. Schultz (Duisburg); he talked about discontinuous optimization arising in stochastic optimization. Dipl.-Math. M. Konik (Chemnitz) gave an approximate description of discrete data from traffic regulation by wavelets. Prof. Dr. E. M. Bednarczuk (Warsaw) examined Hölderian stability of minimal points for cones with possibly empty interior, before Prof. Dr. Sh. Zhang (Hong Kong) demonstrated the virtue of interior point methods for stochastic optimization. It was reserved for Prof. Dr. B. T. Polyak (Moscow) to put the final scientific point by his invited talk on hidden convexity in optimization and control.
Confirmed by positive response of participants, the organizers are very glad and grateful about the two days in Rotterdam. Existing scientific contacts were deepened, new projects initiated. In his concluding words, Prof. Dr. T. Terlaky invited to the following workshop of EUROPT in Edinburgh (2002). I may forward this invitation on you sincerely!
On July 9-17, 2001 the 34th International Workshop of International School of Mathematics "G. Stampacchia", titled "Optimization and Control with Applications" was successfully held in Erice-Sicily at "Ettore Majorana Centre for Scientific Culture". The Workshop was organized by the Department of Applied Mathematics of The Hong Kong Polytechnic University and International School of Mathematics "G. Stampacchia". The Workshop was sponsored by Italian Ministry of University and Scientific Research, Sicilian Regional Government, The Hong Kong Polytechnic University and National Cheng Kung University. The Director of the School is Professor F. Giannessi from University of Pisa. Directors of the workshop are Professor L. Qi and Professor K. L. Teo from The Hong Kong Polytechnic University.
More than 80 researchers from the following countries: Australia, Belgium, Brazil, Canada, Chile, China, Finland, France, Germany, Greece, Hungary, Iran, Italy, Japan, The Netherlands, Pera, Russia, Singapore, South Africa, Spain, Turkey, UK, USA, Venezuela participated in this Workshop. There were many distinguished scientists in Optimization and Control, including Prof. A. Auslender, Prof. J. Burke, Prof. G. Y. Chen, Prof. V. F. Demyanov, Prof. R. Fletcher, Prof. C. Kanzow, Prof. J. M. Martinez, Prof. A. Maugeri, Prof. E. Polak, Prof. T. Rapcsak, Prof. K. Roos, Prof. A. Rubinov, Prof. S. Schaible, Prof. K. Schittkowski, Prof. A. Shapiro, Prof. E. Spedicato, Prof. J. Sun, Prof. M. Todd, Prof. P. Tseng, Prof. S. Y. Wu, Prof. X. S. Zhang.
The Workshop focused on theoretical, modeling and algorithmic issues of Optimization and Control Problems. Recent results in the wide range of the diverse approaches to Optimization and Control Problems were presented and discussed during the nine-day workshop.
The Participants of The workshop enjoyed the informal academic atmosphere in Erice, the town of Science in Sicily, Italy.
Report on the McMaster Optimization Conference: Theory and
Applications
MOPTA 01
August 2-4, 2001, McMaster University, Hamilton, ON
Canada
The "McMaster Optimization Conference: Theory and Applications" (MOPTA 01), hosted by the Department of Computing and Software, McMaster University and its newly established Advanced Optimization Lab, was held successfully on 2-4, August, 2001 in the new Information Technology Center building of McMaster University, Hamilton, Ontario, Canada. The conference attracted about 100 participants coming from different theoretical and applied branches of optimization and provided a great chance for them to interact.
During the conference, seven distinguished researchers gave one-hour invited talks on topic of wide interest. These were: Dimitri Betrtsekas (MIT), John Dennis (Rice University), Ignacio Grossmann (Carnegie Mellon University), Don Jones (General Motor), Micheal Todd (Cornell University), Liven Vandenberghe (University of California, Los Angles) and David Williamson (IBM, Almaden). A total of 43 contributed talks were presented at the conference covering a broad range of theory and applications of optimization, both discrete and continuous.
Details about the conference program, including the transparencies of some speakers, a photo gallery of the events can be found on the confernce WEb page: http://www.cas.mcmaster.ca/~mopta/
The conference was opened by Prof. Mamdouh Shoukri, Vice President Research and the participants were also welcomed by Prof. Mo Elbestavi Dean of Engineering. Both speakers emphasized the importance of optimization in engineering sciences and information technology and that these areas are on McMaster's priority list.
Two special issues (ANOR, OPTE) will be published based on the selected papers from the conference with respect to the theory and applications of optimization.
The conference received generous support from the McMaster Faculty of Engineering, IBM Canada, Materials and Manufacturing Ontario, and The Fields Institute as part of the latter's special year on Numerical and Computational Challenges in Science and Engineering.
The conference committee consists of four professors from McMaster University, S. Kolliopoulos , T. Lou, J. Peng and T. Terlaky (conference chair) and Prof. H. Wolkowicz from Waterloo University.
Confirmed by the success of MOPTA 01, the organizers are looking forward to MOPTA 01 as the closing workshop of Fields Institute's special year on Numerical and Computational Challenges in Science and Engineering.
After about 8 month's work, the new POP web site is ready for use. The address is http://158.132.174.242/pop/. This project was financially aided by the Department of Applied Mathematics, The Hong Kong Polytechnic University, whose support is highly acknowledged.
The structure of the site was the outcome of several meetings among Xinmin Yang, Huoduo Qi, Dehui Li and Xiaoqi Yang. In particular, Huoduo Qi provided an initial list of Research Interests.
The design work was carried by a team (Paula Hodgson (Project Associate), Jeff James, Leo Hung) of the Educational Development Center of the Hong Kong Polytechnic University. They did an excellent job. Their hard work is very much appreciated.
The POP web site is installed in a departmental server of
the Department of Applied Mathematics, at the Hong Kong
Polytechnic University. Joseph Cheng and Joseph Lee have
helped the installation of this package. Their kind help is
appreciative.
The Department of Mathematics at NUS invites applications for several tenure-track and visiting positions in 2002. We will consider outstanding researchers in any field of pure and applied mathematics, particularly those in the areas of Financial Mathematics, Scientific Computing, Optimization and Operations Research, as well as in Computational Biology, Mathematical Modeling, Approximation and Simulations.
Application materials should be sent to
Search Committee
Department of Mathematics
National University of Singapore
2 Science Drive 2, Singapore 117543
Republic of Singapore
Fax: +65 779 5452
and should include: (1) an American Mathematical Society Standard Cover Sheet, (2) a detailed CV including publications list, (3) a statement of research accomplishments and plan, (4) at least three letters of recommendation including one which indicates the candidate's effectiveness and commitment in teaching. Inquiries may be sent via email to search@math.nus.edu.sg.
Review of applications will begin December 15, 2001, and will continue until positions are filled.
Only shortlisted candidates will be notified
International Conference
"Optimization and Industry"
1-6 July, Great Keppel Island, Australia
(http://optimization.cqu.edu.au)
1. Organisers: Professor Panos Pardalos and Dr. Victor Korotkich
2. Organised by: University of Florida,Central
Queensland University, CQU, Center for Applied
Optimization,
University of Florida
3. Endorsed by: Society for Industrial and Applied
Mathematics(SIAM), Activity Group on Optimization,
Australian Mathematical Society, The Pacific
Optimization Research Activity Group(POP)
4. Sponsors: Faculty of Informatics and Communication, CQU, Deal-Fx, ILOG, Mackay Taxi Holdings
5. "Optimization and Industry" Purpose and Results
At the beginning of the new millennium, optimization is
undergoing a major transformation in scope and dimension.
It evolves in response to challenges and opportunities from
industry. From a largely dominant focus on specific problems,
optimization is rapidly expanding to provide all the
ingredients for a surge of a much broader flowering of
technological, business and financial innovations that could
rival any in history.
At the time of this intense change the conference provided a unique opportunity for researches in optimization and captains of industry to understand the significance of this change by reflecting on the fields accomplishments, new developments and future directions. The emphasis of these new frontiers is reflected in the theme for the Conference "Optimization and Industry: Coming together in the New Century".
The Conference gave for people from industry information about latest advances of optimization in telecommunications, finance, power industry, chemical industries, auto-manufacturing, aerospace engineering and computing.
Finding efficient solutions for the world e-economy gives
optimization a potential to enhance immensely its role
for
industry and society as a whole. The Organisers believe that
the Conference was a first attempt to realise the
significance and meaning of this potential for a never-ending
growth of humanity in intelligence and spirit.
6. Conference Plenary Speakers:
"The Australian" wrote about the Conference:
"The massive avalanche of data continually flooding into our
computers is going to need the science of optimization to
control it in the 21 century, according to two scientists
organising a conference on the topic."
"Nowadays, the main task of optimization is to investigate the cutting-edge frontiers of these technologies and systems and find the best solutions for their realisation," said optimization experts Panos Pardalos and Victor Korotkich."
"Morning Bulletin" wrote about the Conference in
the article "Top Scientists Meet"
"A HIGH-powered brains trust meeting in Queensland could lead
to faster aircraft, new drugs and an untold range of new
developments. Comprising 40 of the world's top
mathematicians and computer scientists, and representatives
of multi-national 30 companies, the group is meeting on Great
Keppel Island. Among the global companies at the meeting
which began yesterday were AT&T, Bell Communications,
NASA, General Motors and Boeing Aeronautics.
Professor Panos Pardalos, who heads the Centre for Applied Optimization at Florida University, said one of the ideas under discussion was how to develop a super computer that has brainpower. "We examining the concept of how to give software a brain", Prof. Pardalos said. This could lead to a supercomputer designing items or material not even thought of by human brains. He said topics to be discussed at the Conference included how to make aeroplanes fly faster and safer, revolutionary designs for motor cars, and new medicines and surgical accessories.
Dr. Victor Korotkich of Central Queensland University said: "It is developments that flow from meetings such as this that provide the jobs of the future and improve the world's living standards".
Progress in Optimization: Contributions from
Australasia
edited by Xiaoqi Yang, Alistair Mees, Mike Fisher and Les
Jennings
This book contains 22 chapters which were presented in the fifth Optimization Day (OD) in The University of Western Australia, Perth, Australia, from 29 to 30 June 1998. 'Optimization Day' has been a series of annual mini-conferences in Australia since 1994.
See the 'Contents and Contributors' http://www.wkap.nl/bookcc.htm/0-7923-6286-1
in the book.
Link to the 'Order Page' http://www.wkap.nl/boordfrm.htm?0-7923-6286-1+1
.
Between July 29 and August 2, 2001, the Second International Conference on Nonlinear Analysis and Convex Analysis (NACA2001) was held at Hirosai University in Hirosaki, Japan. This was the first international conference in both areas of Mathematics and optimization, and it had several unique features.
During the last three decades, the study of Nonlinear
Analysis has been devoted vigorously and such activity had
great influence on other areas of science as much as
mathematics. At the same time, Convex Analysis has grown in
connection with the study of problems of optimization,
equilibrium, control, and stability of linear and nonlinear
systems. These two mathematical disciplines have no border
and they rather have good effects each other. Now, we are
requested and asked for new ideas and
human technology based on reliable and theoretical
methodology. We believe that world-wide experts on nonlinear
analysis and convex analysis can answer reasonable and
optimal solutions for such problems if they cooperate with
each other. Based on this idea, we had the first
international conference at Niigata in 1998, and this
NACA2001 was continual and developing more widely.
The program of the conference consisted of invited papers
and contributed papers in eight plenary sessions, six
organized sessions, and regular parallel sessions. Their
parallel sessions were contained in five streams, and all
plenary talks (40 min)
were delivered in the morning sessions and other invited
and/or contributed talks (30 or 40 min) were in the afternoon
sessions.
Totally, there were 129 papers (an increase of 44 papers over
the first conference), arranged into 38 sessions, and 57
people with their families from abroad; totally 150 people
including Japanese. Plenary talks were given by
Professor Anthony To-Ming Lau, University of Alberta
(Canada),
Professor William A.Kirk, University of Iowa (USA),
Professor Masao Fukushima, Kyoto University (Japan),
Professor Hidefumi Kawasaki, Kyushu University (Japan),
Professor Po-Lung Yu, University of Kansas (USA),
Professor Tetsuzo Tanino, Osaka University (Japan),
Professor Hang-Chin Lai, I-Shou University (Taiwan),
Professor Liqun Qi, The Hong Kong Polytechnic University
(China),
Professor Simeon Reich, The Technion-Israel Institute of
Technology (Israel),
Professor Biagio Ricceri, University of Catania (Italy),
Professor Leon A.Petrosjan, St.Petersburg State University
(Russia),
Professor Vladislav I.Zhukovskiy, Russian Institute of
Textile and Light Industries (Russia),
Professor Tsuyoshi Ando, Hokusei Gakuen University
(Japan),
Professor Sehie Park, Seoul National University (Korea),
Professor Wataru Takahashi, Tokyo Institute of Technology
(Japan),
Professor R.Tyrrell Rockafellar, University of Washington
(USA).
The cochairs of the International Program Committee were
Professors Wataru Takahashi of Tokyo Institute of Technology
and Tamaki Tanaka of Niigata University; and the cochairs of
the Local Organizing Committee were Professor Shigehiko
Kuratsubo and Dr.Masamichi Kon of Hirosaki University. The
conference was supported by all participants, in particular
the members of the organizing committee, who have supported
in planning and implementing this conference. Especially, we
wish to thank Dr.Kon of Hirosaki University, who has managed
and arranged all preparation and organization of the
conference.
Also, we are very grateful to Tanaka's graduate students, who
have accepted immediately to take an active part in the
organization and have worked hardly for this conference. Our
gratitude also goes to the various members of the committees
for the time and effort they put into the promotion of the
conference.
Based on these organizations, the participants enjoyed all
programs in the conference including Neputa and Nebuta
Festivals in Hirosaki and Aomori, respectively. Instructions
and other information on submissions for the Proceedings and
Special Issues of
Journal of Nonlinear and Convex Analysis for NACA2001 will be
announced by e-mails to all participants and by the
conference web-page:
http://www.st.hirosaki-u.ac.jp/~naca2001/
**********************************
Interviewee: Lucien Polak
Interviewer: Liqun Qi
Interview Date: September 2001
**********************************
Please answer the following nine questions. After answering
them, please send the questions and the answers via e-mail to
the ORB committee member or the POP member who sent this to
you. If you wish to change some questions, please discuss
with the ORB committee member. The questions and the answers
will appear in the next ORB issue. We mainly send this
questionnaire to some established optimization researchers.
We hope that this may be helpful to young optimization
researchers in this region. In this way, ORB plays a bridge
role between established and young optimization researchers
in this region.
Prof. Elijah Polak
Department of Electrical Engineering
and Computer Sciences
University of California
Berkeley, CA 94720-1770
USA
=================================
email: polak@eecs.berkeley.edu
http://www.eecs.berkeley.edu/~polak
phone: 510-841-4546
FAX: 510-841-4546
=================================
Ph.D, University of California, Berkeley, 1961
Over 250
Nonsmooth optimization, optimal control, semi-infinite
optimization, optimal engineering design.
E. Polak, "Computational Methods in Optimization: A Unified Approach", Academic Press, 329 pages, 1971.
E. Polak, "Optimization: Algorithms and Consistent Approximations", Springer, New York, 800 pages, 1997
Polak, R. Trahan and D. Q. Mayne, "Combined phase I -
phase II methods of feasible directions", Mathematical
Programming,
Vol. 17, No. 1, pp. 32-61, 1979.
C. Gonzaga and E. Polak, "On Constraint Dropping Schemes and Optimality Functions for a Class of Outer Approximations Algorithms", SIAM J. Control and Optimization, Vol. 17, No. 4, pp. 477-493, 1979.
L. He and E. Polak, "An Optimal Diagonalization Strategy for the Solution of a Class of Optimal design Problems", IEEE, Trans. on Automatic Contr., Vol. 35, No.3, pp 258-267, 1990.
E. Polak, "On the Use of Consistent Approximations in the
Solution of Semi-Infinite Optimization and Optimal Control
Problems", Math Programming, Series B, Vol. 62, No.2, pp
385-414, 1993.
The Polak-Ribiere method of conjugate directions.
The development of the concepts of "conceptual" and "implementable" algorithms, together with a theory of optimization algorithms based on algorithm maps and providing general conditions for the convergence of algorithms.
The development of a number of nonlinear programming,
min-max, semi-infinite optimization, and optimal control
algorithms.
| NAME | YEAR OF PH.D | ORIGIN | |||||||||
| B. W. Jordan | 1964 | US | |||||||||
| M. D. Canon | 1965 | US | |||||||||
| C. D. Cullum | 1965 | US | |||||||||
| A. Larsen, Jr. | 1965 | US | |||||||||
| J. P. Jacob | 1966 | Brazil | |||||||||
| N. O. Da Cunha | 1966 | Brazil | |||||||||
| K. Y. Wong | 1966 | HK | |||||||||
| E. J. Messerli | 1968 | Canada | |||||||||
| G. L. Meyer | 1969 | France | |||||||||
| O. Pironneau | 1971 | France | |||||||||
| R. Klessig | 1972 | US | |||||||||
| H. Mukai | 1974 | Japan | |||||||||
| I. Teodoru | 1975 | Romania | |||||||||
| L. J. Williamson | 1975 | US | |||||||||
| R. Trahan | 1976 | US | |||||||||
| A. N. Payne | 1978 | US | |||||||||
| S. Tishyadhigama | 1978 | Thailand | |||||||||
| A. Tits | 1979 | Belgium | |||||||||
| Y. Wardi | 1981 | Israel | |||||||||
| D. M. Stimler | 1984 | Australia | |||||||||
| T. L. Wuu | 1986 | Taiwan | |||||||||
| S. E. Salcudean | 1986 | Canada | |||||||||
| T. E. Baker | 1988 | US | |||||||||
| J. E. Higgins | 1989 | Ireland | |||||||||
| Y-P. Harn | 1989 | Taiwan | |||||||||
| L. He | 1990 | PRC | |||||||||
| J. Wiest | 1990 | US | |||||||||
| T. H. Yang | 1991 | US | |||||||||
| A. L. Schwartz | 1996 | US | |||||||||
| C. Kirjner-Neto | 1996 | Brazil | |||||||||
| I. Khalil | 1997 | Iraq |
The development of a theory of consistent approximations
for the solution of optimal control, shape optimization, and
semi-infinite optimization problems.
The development of optimization algorithms for problems in which constraint or cost functions are of the form
f(x) = max_{y \in Y} min _{z \in Z}\phi(x,y,z) These occur in design centering, tolerancing and tuning engineering problems,
f(x) = \int_{y \in \Omega(x)} g(x,y) dy these occur in engineering optimization problems with reliability cost or constraint,
f(x) = max_{y \in Y(x)} \phi(x,y) these are of intrinsic mathematical interest, but they also occur as a by product in engineering design.
Lucien Polak spent his 70th birthday in Sydney on August 11, 2001. Taking this occasion, we conducted an ORB Personal Interview with him. His answers to the ORB questionnaire are in the ORB article before this article.
Lucien was originally from Poland. (His last name "Polak"
means "from Poland".) He is a a survivor of the Nazi
Holocaust. After the World War II, only he and his mother
in his family survived the Nazi Holocaust. They immigrated
to Australia without any money. Lucien even missed regular
high school education. But he survived all of these. He
entered University of Melbourne, obtained his B.E. there.
Then he went to the USA. He obtained his Ph.D. degree in
University of California,
Berkeley in 1961, and eventually became a professor in that
world-class university.
Lucien made contributions in both optimization and optimal control. In unconstrained optimization, the Polak-Ribiere method of conjugate directions is very well-known. In minimax optimization, semi-infinite optimization, nonsmooth optimization, optimal control, Lucien developed a number of algorithms which are quite influential.
Lucien has published over 250 papers, including two
well-known books: "Computational Methods in Optimization: A
Unified
Approach" (1971) and "Optimization: Algorithms and Consistent
Approximations" (1997). More than 30 persons obtained Ph.D.
degrees under his supervision. Some of them are well-known
themselves now.
Lucien is is a friend of us. He is always friendly and helpful to others. We wish that he will continue his research in optimization and control for many years ahead.
For further details, please click here.
Lucien Polak received the B.Sc. degree in Electrical Engineering from the University of Melbourne, Australia, in 1957, and the M.S. and Ph.D. degrees in Electrical Engineering from the University of California at Berkeley in 1959 and 1961, respectively. Since 1961 he has been on the faculty of the University of California at Berkeley, where he is presently a Professor Emeritus of Electrical Engineering and Computer Sciences and a Professor in the Graduate School.
Lucien has been the leading figure in the field of continuous-parameter optimization. He made pioneering contributions in diverse areas ranging from conjugate-gradient algorithms and constrained optimization to semi-infinite programming and min-max problems, to nondifferentiable optimization and large-scale, infinite-dimensional problems. In all of these areas he has been in the forefront, extending the state of the art and defining new research directions. What distinguishes his research and puts it on the highest plateau is his dual approach to optimization theory and practice.
On the theoretical side, Lucien has developed a conceptual framework for algorithms, based in part on algorithm models with general, provably-convergent properties. An algorithm having one of the prototype models can first be presented in a simplified conceptual form, and then further developed into an implementable form. Its convergence properties are evident in light of the underlying model, while its implementable version can be tailored to specific optimization problems. This general paradigm greatly simplifies the task of developing provably-convergent algorithms for large classes of problems, and has provided rigorous mathematical foundations to the art of algorithms development.
On the practical side, Lucien has pioneered a systematic, optimization-based approach to engineering systems design, and applied it in various technological areas such as control systems, circuits and structures. The associated optimization problems often are large scale, and involving either infinite-dimensional parameter spaces or infinitely many constraints. They often are cast in the setting of optimal control, semi-infinite programming, or systems described by partial differential equations. Such problems require infinite time for function evaluations, and hence are compounded by computational challenges. To address this difficulty, Lucien developed the theory of consistent approximations, providing the theoretical foundations to a suite of adaptive- precision algorithms that find an appropriate balance between the conflicting objectives of numerical precision and low computational complexity. He had used such algorithms for solving optimal control problems and structural optimization problems, and applied similar ideas (concerning adaptive precision) to constraint-dropping schemes for semi-infinite optimization.
Lucien supervised thirty-one Ph.D. students. Many of us are currently working in far-apart disciplines ranging from mathematics to circuit design and telecommunications. But we all share a common element in our professional makeup that is part of Lucien’s legacy. I have gone far afield to an area that has little to do with optimization, but I often find the clue to challenging problems in the strong foundations that Lucien imparted to his students.
I spent two years working under Lucien’s supervision, and this has been an extremely important and wonderful period in my life. Besides control and optimization I learned from him a basic approach to science. Like many of his students I got to know him on a personal level through trekking the Sierras and hiking along the Pacific shore. At this opportunity of celebrating your seventieth birthday, I thank you Lucien for a life-shaping experience and I wish you many more fruitful years.
I distinctly remember my first meeting with Lucien, the day after I arrived in Berkeley, in early September 1977. I knew about his international fame in the field of numerical optimization, as I had consulted his 1971 book and several papers of his, as well as papers referring to his work, in the course of my bachelor's thesis work at the University of Li\`ege. When I had received my admission letter from Berkeley, which at the same time informed me that E. Polak would be my academic advisor, I was thrilled beyond imagination. And now I was going to meet him in person. I must say that I was a bit nervous. But when I entered his office that day, the world famous professor immediately made me feel at ease. We talked about my long journey, my first few hours in California---I was staying at the International House, I had already walked about town and up in the hills---, my family, my extra-curricular interests. One thing that sticks in my memory is that he cautioned me about sharks if I was going to swim in the ocean... Behind the eminent scholar I had just discovered a warm human being, with a sharp sense of humor. This meeting was to mark the start of four years of close interaction with Lucien, as well as the start of a lifetime friendship.
In June 1978---as soon as I got through the daunting
"prelims"--Lucien offered me a research assistantship
position, which I immediately accepted. For the next three
years, till I left Berkeley in August 1981, I spent an
incredible number of hours working with him, mostly in his
office but also, many a Sunday afternoon, in his house, up in
the hills. While he did have one or two other students at
the time, and of course many other responsibilities besides
mentoring his research advisees, Lucien was
available to meet with me whenever I needed his advice, or so
it seemed. In such privileged setting I learned enormously
from him. Uncountable times, when I was at loss about how to
modify a tentative algorithm to obtain certain desirable
properties, he gave me enlightening geometric intuition,
which often would render the issue all but trivial.
Uncountable times, when I was unsuccessful in proving certain
convergence properties, he steered me in a new, promising
direction that I had not thought about. Over time, by
osmosis, I slowly acquired some of these skills for myself.
Above all, I inherited from Lucien a taste for uncompromising
rigor and precision.
Our interaction during my Berkeley years went way beyond research and classes. In particular, I had the privilege to accompany him on several backpacking trips in Yosemite National Park. This included one cross-country skiing trip, complete with snow camping in the wilderness. This was a totally new experience for me. There again, I learned a lot from Lucien.
After I left Berkeley, we stopped writing papers together, but we remained friends and stayed in close contact with each other. Most importantly, the research qualities I learned from Lucien will always remain with me. I have also thrived to the best of my abilities to pass them on to my own students at the University of Maryland.
To me, Lucien Polak was the best research advisor I could have dreamed of. From the bottom of my heart, thank you Lucien.
This is the photo of Lucien at the 15th Math Prog Symposium, Ann Arbor, 1994. From left to right: Roman Polyak, Lucien Polak and Boris Polak. To see the photo, please click: ThePol.jpg . We are not relatives, but work in one field!
("Polak" and "Polyak" are the same family name in Polish, which means "from Poland". This is like "Chang", "Teo" and "Zhang" are the same Chinese family name, with the same Chinese character. Now, Lucien and Roman are in the USA, while Boris is in Russia. They have made their names known in the optimization community by their own outstanding achievements. --- ORB editor)
