Cursos

• CURSO 1: Dr. Jay Boisseau (San Diego Supercomputer Center, University of California, San Diego, USA): Developing High Performance Applications for Parallel Computers Parallel computing has become the dominant technique for achieving high performance in computational science and engineering research. Parallel computing systems are now becoming mainstream in commercial sectors as well, due to the performance demands and requirements of today's engineering, database, and financial applications. Multiprocessor systems based on commodity processors (IA32, G4, Alpha, UltraSparc, and Power3) are now common and offer excellent performance for the price. However, formal educational opportunities for learning parallel computing are still relatively rare. Parallel computing often must be learned `on the side' (for research scientists) or `on the job' (in industry). This course, based on a university-level class the instructor has taught at San Diego State University and the University of California San Diego, is an intensive introduction to parallel computing for scientists and engineers. The class starts with a high-level overview of parallel computing, including parallel algorithms, Amdahl's Law, and limits to scalability. It will then compare and contrast the architectural features of modern parallel computers, including commodity clusters. The most recent programming models and tools for these systems will then be presented. Finally, the content will discuss effective methods for developing, optimizing, and debugging scientific applications for parallel computers. The course will be very ``applied'' in nature. The lectures will place more emphasis on developing parallel applications than on abstract theoretical parallel programming concepts or computer engineering details. The labs will reinforce this practical focus by letting students work with real parallel applications and develop new applications from (small) serial codes. Attendees will leave this course ready to apply parallel computing expertise to develop new, high-performance applications for their research activities.
• CURSO 2: Dr. Michael Collins (Naval Research Laboratory, Washington, DC, USA): Parabolic Equation Techniques with Applications This course will cover several types of nonseparable wave prophagation problems in the geosciences, including ocean acoustics, seismology, athmospheric waves, and waves in porous media. Prof. William Siegmann (Department of Mathematics Sciences, Rensselaer Polytechnic Institute) and Dr. Collins are presently writing a book on this topic that will be completed in time to use in the short course. Computer laboratory sections would be very convenient for this course.
• CURSO 3: Dr. Gene H. Golub (Computer Science Department, Stanford University, USA): Methods for Computing Eigenvalues This course will discuss modern methods for computing eigenvalues/vectors of matrices. Particularly in solving for a few eigenvalues of large, structured matrices, symmetric and general. It will describe specialized techniques for solving the quadratic eigenvalue problem and modified eigenvalue problems. In addition, methods for bounding quadratic and bilinear forms using eigenvlue analysis will be presented.
• CURSO 4 : Dr. Bertil Gustafsson (Department of Scientific Computing, Uppsala, Sweden): Hyperbolic Systems and Numerical Methods. Problems with periodic solutions. First order systems in one space dimension. First order systems in several space dimensions. Difference approximations with constant coefficients. Difference approximations with variable coefficients. The method of lines. The finite volume method. The Fourier pseudospectral method. Initial-boundary value problems. The energy method for hyperbolic systems. The Laplace transform method. The energy method for difference approximations. The Laplace transform method for difference approximations. The method of lines and generalized stability.
• CURSO 5: Dr. Cristina Pereyra (Department of Mathematics and Statistics, University of New Mexico, Albuquerque, USA): Wavelets: Applications and Theory Wavelets have become part of the toolbox of scientist. Wherever a signal or image needs to be analyzed, the wavelet transform can be used. Wavelets provide a ``mathematical zoom" that permits one to analyze functions and operators at many scales simultaneously. Wavelets have wonderful approximation properties. From the mathematical point of view they provide bases for a number of classical spaces of functions. From the practical point of view they permit to represent certain signals very efficiently. Wavelets are being used to study turbulence and PDE's. They also have become a popular denoising tool. In this course, Dr. Pereyra will explain: (a) the basics of wavelets and signal/image compression; (b) how to construct divergence-free multi-wavelets which are expected to be useful in numerical analysis of incompressible fluids; (c) denoising procedures and compare wavelet techniques to classical denoising techniques; and (d) what wavelets are, how to construct them, and how to implement them.
• CURSO 6: Cristina Maciel (Department of Mathematics, Universidad Nacional del Sur, Bahía Blanca, Argentina): The Trust Region Strategy for Nonlinear Optimization Problem Ten lectures of 2 hours each, including laboratory's experiments by using MATLAB. The purpose of this course is to illustrate and describe the role of trust region strategy in Nonlinear Optimization. The discussion in the first half of the course will be devoted to the trust region algorithms for the unconstrained optimization and nonlinear systems of algebraic equations. In the last half of the course, after a review of the sequential quadratic programming methods for the general nonlinear programming problem, the trust region strategy for the constrained optimization problem is presented. This is a rich and deep topic requiring much more than four lectures to cover with any rigor. Nevertheless, this topic will be introduced highlighting issues that involve optimization. The course concludes by presenting recent developments in algorithms and software for trust region approach. Attendees will obtain an understanding of state-of-the-art in trust region methods for nonlinear optimization problems. Contents: The trust region strategy for the unconstrained optimization problem. The problem. Description of the general algorithm. The trust region subproblem. Different strategies for solving the trust region subproblem. The large scale case. Global convergence theories. The trust region strategy for the nonlinear systems. The trust region strategy for the general constrained optimization problem. Description of the general algorithm. The trust region subproblem. The tangent space and the full space approaches. Merit functions. Global convergence theories. Recent developments for trust region methods.
• CURSO 7: Reinaldo González (Escuela de Ingeniería de Petróleo, Facultad de Ciencias, Universidad Central de Venezuela) : Petroleum Reservoir Simulations Analysis of Pressure Transient for Horizontal Wells with Irregular Trajectory in Arbitrary Shaped Reservoirs.
• CURSO 8: Juan Meza (Computational Sciences and Mathematics Research Group, Sandia National Laboratories, Livermore, CA, USA): Pattern Search methods for Optimization The advent of powerful desktop computers and small to medium scale shared memory multiprocessors (SMP) has dramatically increased the use of optimization in engineering and scientific disciplines. This is especially evident in simulation-based problems where the objective function is represented by a computer model of a physical or engineering process. Within the optimization community, this trend has spawned new research areas as well as revitalized some old areas. One good example is in the area of parallel optimization. It has been proposed that there are three major levels for introducing parallelism in optimization problems: 1) parallelize the function, gradient, and constraints, 2) parallelize the linear algebra, and 3) parallelize the optimization at a high level. The third option, that of parallelizing optimization at a high level, will be one of the main focus areas covered in this course. The course is designed both for students wishing to understand the optimization aspects as well as people who are already practicing optimization in their work or research. This course will cover the basic mathematical concepts of optimization, followed by material on optimization model building and problem classification. Typical areas covered from the practical side include, how to recognize when one has a solution, difficulties when solving optimization problems, and setting up and handling constraints. The majority of the lectures will be focused on optimization methods aimed for parallel computing and for problems where the objective function is based on the solution of a simulation. Examples of methods that will be discussed include parallel direct search methods, genetic algorithms, and simulated annealing. We will also present some new hybrid methods that combine pattern search methods with trust-region methods and suggest new areas of research.

Comité Organizador Local

Empresas y Organismos Auspiciantes

* IX ELAVIO (IX Escuela Latinoamericana de Verano de Investigación Operativa)

La IX ELAVIO se llevará a cabo del 25 de febrero al 1 de marzo de 2002 en Vaquerías, Provincia de Córdoba, Argentina organizada por el Departamento de Computación de la Facultad de Ciencias Exactas y Naturales de la Universidad de Buenos Aires, con la colaboración de profesores de otras universidades nacionales y profesores de otros países latinoamericanos. Las ELAVIO están dirigidas principalmente a investigadores jóvenes y estudiantes de doctorado o de grado avanzado de países latinoamericanos. Los objetivos principales de las ELAVIO son:

• brindar un panorama actualizado en algunos temas de punta a investigadores recién formados o en formación de la región,
• promover el intercambio científico y tecnológico entre investigadores senior y jóvenes de las áreas de Optimización e Investigación Operativa.
• fomentar el conocimiento mutuo, que derive en trabajo en colaboración entre los participantes.

* ICMI-EARCOME 2002 / SEACME 9

Este Congreso se desarrollará durante el mes de mayo de 2002, en el National Institute of Education, Nanyang Technological University, Singapore. Es organizado por el International Programme Committee. Más información sobre el mismo se puede encontrar en el Noticiero nro. 28.

c/o Conference and Travel Management Associates Pte Ltd, 425A Race Course Road, Singapore 218671, Republic of Singapore.

Website: http://math.nie.edu.sg/earcome/

* 24th World Conference on Boundary Element Methods Incorporating Meshless Solutions Seminar

Esta Reunión tendrá lugar en Sintra, Portugal, entre el 17 y 19 de Junio de 2002. Es organizada por el Wessex Institute of Technology, UK y la Universidad de Coimbra, Portugal. Es auspiciado por la International Society of Boundary Elements (ISBE) y el International Journal of Engineering Analysis with Boundary Elements. Más información sobre este Congreso se encuentra en el Noticiero nro. 28. Submission of abstracts:

• electronically either through the website or via E-mail:
• shanley@wessex.ac.uk
• Fax: 44 (0) 238 029 2853.
• Informes: Susan Hanley, Conference Secretariat, BEM 24, Wessex Institute of Technology, Ashurst Lodge, Ashurst, Southampton, SO40 7AA, UK.

* Non Linear Analysis, Function Spaces and Applications (NAFSA 7)

Se desarrollará en Praga, República Checa, entre el 17 y 22 de julio de 2002, organizado por la Academy of Sciences of the Czech Republic y el Department of Mathematics of the Czech University of Agriculture. Más información sobre este Congreso se encuentra en

Informes: NAFSA7, Lubos Pick, Katedra matematicke analyzy, Matematicko-fyzikalni fakulta UK, Sokolovska 83, 186 75 Praha 8, Czech Republic.

E-mail: nafsa7@karlin.mff.cuni.cz

* Primer Evento Internacional de Matemática para Ingeniería

Este Evento es organizado por el Departamento de Matemática Aplicada de la Facultad de Ingeniería Eléctrica, Universidad de Oriente, Santiago de Cuba. Se desarrollará entre el 12 y el 14 de julio de 2002.

Fax: 53 01 226 653539,         E-mail: mat@fie.uo.edu.cu

* VIII Encuentro Rioplatense de Álgebra y Geometría Algebraica

Durante los dias 29 y 30 de Noviembre de 2001 se realizó en la Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, el VIII Encuentro Rioplatense de Algebra y Geometría Algebraica. Se desarrolló el siguiente programa de actividades:

• Beatriz Abadie, Universidad de la República (Montevideo): ``Sobre C*-álgebras"
• Marco Farinati, Universidad de Buenos Aires: ``Sobre extensiones de Galois de anillos"
• Cristian González Aviles, Universidad de Chile: ``Motivos de variedades algebraicas"

Conferencias

- Andrés Abella, Universidad de la República (Montevideo): ``Categorías trenzadas"

Comité Organizador: Andrés Abella (andres@cmat.edu.uy); Fernando Cukierman (fcukier@mate.dm.uba.ar); Walter Ferrer (wrferrer@cmat.edu.uy); Andrea Solotar (asolotar@mate.dm.uba.ar).

* Congrès de mathématiques appliquées à la mémoire de Jacques-Louis Lions

Place: Collège de France, Paris, July 1-5, 2002

Speakers: Luigi Ambrosio, François Baccelli, John Ball, Franco Brezzi, Luis Caffarelli, Marie-Paule Cani, Alexandre Chorin, Jean-Michel Coron, Lawrence Evans, Olivier Faugeras, Mathias Fink, Michael Ghil, Thomas Hou, Andrew Majda, Louis Nirenberg, George Papanicolaou, Anthony Patera, Benoît Perthame, Rolf Rannacher, Panagiotis Souganidis, Eitan Tadmor, Srinivasa Varadhan, Cédric Villani, Mark Vishik, Jean-Christophe Yoccoz, Enrique Zuazua.

Honour Committee: Hubert Curien, Hiroshi Fujita, Peter Lax, Enrico Magenes (Chair), Guri Marchuk.

Patronages: International Mathematical Union and Académie des Sciences de Paris.

Supported and sponsored by: Ministère de la Recherche, CNRS, CNES, INRIA, Collège de France, École Polytechnique, Université Pierre et Marie Curie (Paris VI), SMAI and SMF.

Registration: Registration to the Conference is requested and free of charge. Please register directly on the web site.

Contacts: Congrès de mathématiques appliquées à la mémoire de Jacques-Louis Lions.