Subject: NA Digest, V. 97, # 17 NA Digest Sunday, April 27, 1997 Volume 97 : Issue 17 Today's Editor: Cleve Moler The MathWorks, Inc. moler@mathworks.com Today's Topics: Seeking Reference on Combustion Problem New Book on Linear Optimization FTP Site for Numerical Linear Algebra Textbook Multi-grid Methods for Biharmonic Equation Updated AMPL/Solver Interface Intellectual Property Rights Change of Address for Gavin Pringle Call for Nominations, DiPrima Prize Call for Nominations, Polya Prize Interface of Computing Science and Statistics Position at New South Wales Centre Position at UCLA Position at Humboldt-Unversity Berlin Studentship at University of Manchester Postdoctoral Position at Bell Laboratories Research Faculty Position at Rice University Position at University of Graz Studentship at University of Bath Postdoctoral Position at University of Bath Positions at EPFL, Lausanne, Switzerland Contemts, IEEE Computational Science & Engineering Contents, Computational Geosciences Contents. Optimization Methods and Software Submissions for NA Digest: Mail to na.digest@na-net.ornl.gov. Information about NA-NET: Mail to na.help@na-net.ornl.gov. URL for the World Wide Web: http://www.netlib.org/na-net/na_home.html ------------------------------------------------------- From: Willy Govaerts Date: Mon, 21 Apr 1997 08:06:42 +0200 (MET DST) Subject: Seeking Reference on Combustion Problem There is a simple combustion problem that can be modelled as \frac{\partial u}{\partial t}= \frac{d^2u}{dx^2} /a^2 + [{1-x}/{2u} - (1-1/u)] [{1+x}/{2u}-(1-1/u)] e^{\theta(1-\frac{1}{u})}}/{\theta } It describes the steady - state combustion of two gases meeting in a gap between porous walls a distance $2a$ apart. Fuel diffuses in at one wall, oxidant at the other and there is a zone of reaction in the middle, dying out towards each wall. Boundary conditions are u(\pm 1)= 1. Here x is the scaled distance from one of the walls and u is the scaled temperature. Finally, $\theta$ is a physical property of the two gases involved. Compared to other combustion models this one is very simple. So it is rather interesting for paedagogical purposes. Unfortunately I do not have any published references. Can one of the NA - netters help me? Many thanks in advance, Willy Govaerts Department of Applied Mathematics and Computer Science University of Gent Krijgslaan 281 - S9 B - 9000 Gent Belgium ------------------------------ From: C. Roo Date: Mon, 21 Apr 1997 18:35:36 +0200 Subject: New Book on Linear Optimization We would like to announce the availability of the book C. Roos, T. Terlaky and J.-Ph. Vial. Theory and Algorithms for Linear Optimization: An Interior Point Approach, John Wiley, Chichester, 1997. More information is available at the URL: http://ssor.twi.tudelft.nl/Publications/Books/TAfLOAIPM/ipmbook.html Best regards, Kees ------------------------------ From: Biswa Datta Date: Mon, 21 Apr 97 14:20:39 CDT Subject: FTP Site for Numerical Linear Algebra Textbook FTP SITE FOR MATCOM The software package MATCOM that implements the major algorithms in my book : Numerical Linear Algebra and Applications, Brooks/Cole publishing Company, 1995 has been revised, updated, and put in the MathWorks anonymous FTP site at ftp.mathworks.com in the directory /pub/books/datta. Please also note that the book has been recently reprinted. I strongly suggest that the copies of the book from the latest print ( Third Print , published in July,1996) be used in the class rooms. Biswa Nath Datta (dattab@math.niu.edu) ------------------------------ From: Matthew Cordery Date: Wed, 23 Apr 1997 15:07:50 -0700 (PDT) Subject: Multi-grid Methods for Biharmonic Equation I am interested in multigrid methods for solving the biharmonic equation on unstructured 2D triangular meshes and am wondering if anyone has any experience in this problem that they might be willing to share. In particular, I am interested in solutions to the biharmonic equation that arises from the equations for creeping flow (Stoke's equations). The fluid itself has a strongly temperature-dependent viscosity that may vary sharply over short distances (relative to the size of the compuational domain). Thus, my biharmonic equation would have a viscosity term embedded within it. Thanks in advance for any help! Dr. Matthew J. Cordery cordery1@llnl.gov Environmental Computer Applications Lawrence Livermore National Laboratory L206 P.O. Box 808 Livermore, CA 94550 ------------------------------ From: David Gay Date: Fri, 25 Apr 1997 22:13:08 +0400 Subject: Updated AMPL/Solver Interface Now (finally!) available is a reorganized AMPL/solver interface library that offers significant enhancements for support of nonlinear solvers: -- Detection of partially separable structure: automatic recognition of a nonlinear function as the sum of terms that depend on only of few variables each (possibly after a term-dependent linear change of variables). -- Computation of 2nd derivatives: use of the partially separable structure to efficiently compute Hessian-vector products and explicit Hessians (of the Lagrangian function in a nonlinear programming problem). New utilities make interfacing with solvers -- both linear and nonlinear -- easier. An expanded set of example interfaces includes a long-promised interface with LANCELOT, as well as a MATLAB interface. AMPL is a language and modeling environment for expressing and manipulating linear and nonlinear programming problems. Much more information is available from the AMPL web site, http://www.ampl.com/ampl/ which offers both a Postscript file, ftp://netlib.bell-labs.com/ampl/REFS/hooking2.ps.gz and a corresponding HTML file, http://www.ampl.com/ampl/REFS/HOOKING/ that describe the updated solver interface library and give pointers to its source. ------------------------------ From: Nicolas Robidoux Date: Sat, 26 Apr 1997 14:55:46 -0600 Subject: Intellectual Property Rights In NA Digest, V. 97, # 12, I posted a query about --what rights Universities and Colleges have or appropriate themselves regarding software which is brought to the institution by a new hire, who subsequently turns it into a commercial product; --in general, what to expect as far as property rights of academic institutions, in particular as pertaining to developing commercial software, and what to watch for in this regard. The following is taken from a summary of my impression of the situation based on the replies of Grzesiek Adamiec, Brian Borchers, George F. Corliss, Henry Greenside, Steve Hammond, Russ Merris, Jonathan R. Shewchuk, Gio Wiederhold and someone wishing to remain anonymous, all of whom I thank. The unabridged version, including pointers to additional information, is found in the newsletter Concerns of Young Mathematicians, Volume 5, Issue 11 dated 23 April 1997, available on the web at the YMN Newletter Archive http://www.math.usouthal.edu/ymn.html and by anonymous ftp to ftp.ms.uky.edu in pub3/mailing.lists/ymn-list. The big picture is that: ++All may be negotiable. Consult with a knowledgeable lawyer. ++The rights of academic institutions over software developed while one is at their employ vary greatly from institution to institution, and within institutions from job title to job title. In some cases they extend to software developed by students. ++A rule of thumb is that anything developed on your own time using your own equipment (as opposed to "university owned or maintained") is your intellectual property. Keeping the development work off the institution's computer network is a good idea. So is keeping it off the institution's office space. ++Nearly all universities have partial ownership of software developed specifically under an academic grant, but many share ownership with the inventor, and don't claim ownership of software that wasn't developed with the aid of grant money or significant university resources. A few universities have policies assigning themselves full ownership and royalties for anything their employees do. ++The situation is complicated by the fact that policies seem to be in flux: PREPARE FOR TURNS FOR THE WORSE. ++The best time to negotiate and record exceptions to institutional rules (including possible future changes in policy) is at hire time, before signing an employment contract. Utter formality in these matters is often unnecessary, but there needs to be a paper trail. ++As far as previously developed software goes, things are equally variable. Have it recorded (in writing, of course) as early as possible (preferably at hire time) that your preexistent software products Widget and Gadget are off limit to the institution no matter what, or under conditions satisfactory to you. Prior work should be disclosed before employment, to protect it from claims. If the work is a substantial reason for the hire, and will be continued, then a more specific contract is advisable. ++Watch what you sign away! Be even more careful with industry. Nicolas Robidoux Albuquerque Resource Center of the High Performance Computing, Education and Research Center. mia@math.unm.edu ------------------------------ From: Gavin Pringle Date: Tue, 22 Apr 1997 13:06:11 +0100 Subject: Change of Address for Gavin Pringle I have recently changed my employer, but not my home address. My new location is Edinburgh Parallel Computing Centre Edinburgh University James Clerk Maxwell Building Mayfield Road, Edinburgh EH9 3JZ, UK. Tel. +44 (0)131 650 6709 Fax. +44 (0)131 650 6555 email: gavin@epcc.ed.ac.uk Cheers, Gavin ------------------------------ From: Allison Bogardo Date: Tue, 22 Apr 97 14:18:55 EST Subject: Call for Nominations, DiPrima Prize CALL FOR NOMINATIONS for THE RICHARD C. DiPRIMA PRIZE The DiPrima Prize SIAM will present the award at the 1998 SIAM Annual Meeting in Toronto, Canada, July 13-17. The award honors the memory of Richard C. DiPrima, long-time Chair of the Department of Mathematical Sciences at Rensselaer Polytechnic Institute and past-president and energetic supporter of SIAM. The award will be based on an outstanding doctoral dissertation in applied mathematics. Eligibility The award, based on Ph.D. research in applied mathematics (defined as those topics covered in SIAM journals or series) is made to a young scientist. The Ph.D. thesis and all other Ph.D. requirements should have been completed in the time period from July 1, 1995 to June 30, 1997. The Ph.D. degree must be awarded by December 31, 1997. Description of the Award The award will consist of a certificate and a cash prize of $1,000. The SIAM president will notify the recipient of the award in advance of the award date and invite the recipient to attend the annual meeting to receive the award. Nominations Nominations, along with a copy of the dissertation (in English), should be sent by November 30, 1997 to: Professor Gilbert Strang Chair, DiPrima Prize Selection Committee c/o Allison Bogardo SIAM 3600 University City Science Center Philadelphia, PA 19104-2688 Telephone: (215) 382-9800 Fax: (215) 386-7999 E-mail: bogardo@siam.org Members of the selection committee are Philip Holmes (Princeton University), Gilbert Strang (MIT), and Shmuel Winograd (IBM Research Center). ------------------------------ From: Allison Bogardo Date: Tue, 22 Apr 97 14:47:52 EST Subject: Call for Nominations, Polya Prize CALL FOR NOMINATIONS for GEORGE POLYA PRIZE The Polya Prize SIAM will present the award at the 1998 SIAM Annual Meeting in Toronto, Canada, July 13-17. The award honors the memory of George Polya and will be given for a notable contribution in one of the following areas: approximation theory, complex analysis, number theory, orthogonal polynomials, probability theory, or mathematical discovery and learning. Eligibility There are no restrictions except that the prize is broadly intended to recognize specific work. Description of Award The award will consist of an engraved medal and a $20,000 cash prize. Nominations A letter of nomination, including a description of achievement(s), should be sent by October 1, 1997, to: Professor Harry Kesten Chair, Polya Prize Selection Committee c/o Allison Bogardo SIAM 3600 University City Science Center Philadelphia, PA 19104-2688 Telephone: (215) 382-9800 Fax: (215) 386-7999 E-mail: bogardo@siam.org Other members of the selection committee are Lennart Carleson (Royal Institute of Technology, Stockholm), Barry Mazur (Harvard University), Paul Nevai (The Ohio State University), and Andrew Yao (Princeton University). ------------------------------ From: David Scott Date: Mon, 21 Apr 1997 11:24:20 -0400 Subject: Interface of Computing Science and Statistics Final Announcement: INTERFACE '97 1997 Themes: Data Mining, Massive Data Sets, Environmental Science The 1997 Symposium on the Interface of Computing Science and Statistics will be held May 14-17, 1997 at the Houston Medical Center Holiday Inn Hotel in Houston, Texas. The Symposium is being organized around the theme of "Mining and Modeling Massive Data Sets In Science, Engineering, and Business", with a sub-theme of multimedia education and quantitative environmental science. The conference is sponsored by the Interface Foundation of North America, a non-profit educational corporation. The Statistics Department at Rice University is hosting the meeting with Dr. David W. Scott as chair. Dr. Jerry Friedman from Stanford University will give the keynote address. The invited program includes some thirty sessions on topics including virtual reality, marketing applications, multimedia education, numerical methods, pattern recognition, visualization, mapping, environmental statistics, wavelets, dimension reduction, computational biochemistry, Bayesian methods, networks and clusters of workstations, virtual departments, and information retrieval for massive data sets. The list of invited speakers includes representatives from academia (Rice, CMU, UW, Baylor, others), industry (IBM, SGI, SAS, Bellcore, others) and government (USDA, JPL, Census, DOD, others) and includes a substantial number of international speakers and participants (Berlin, Glasgow, Belgium, Finland, others). For many years advances in statistics and particularly statistical computation have been driven by the general demands of industry. The explosion of on-line resources and performance-price computer power has dramatically increased expectations of what computational and statistical scientists can provide. Data warehousing is a new trend in industry which is intended to provide information support to all segments of a business. Government is moving to provide on-line access to many of its databases. These databases are approaching terabyte size. This explosive growth is being matched in many academic research labs, libraries, among others. The problems of efficiently and effectively searching and modeling based on such massive data sets are the focus of the 1997 Interface symposium. Many traditional statistical and computational tools have been brought to bear on these types of problems. Innovative visualization methods, where possible, often shed insight on underlying structure. Novel computational paradigms, such as neural nets and Bayesian methods and artificial intelligence tools, are required in order to even begin to understand and model the data. There are a wealth of ad-hoc techniques developed by both statisticians and computational scientists. An active discussion, between subject matter specialists and statistical and computational practitioners is sure to lead to a fruitful interchange. Our intention is to foster such a dialog via the Interface conference. For further information, please contact: Dr. David W. Scott EMAIL: interface97@stat.rice.edu TELEPHONE: (713) 527-6037 FAX: (713) 285-5476 WEB URL: http://www.stat.rice.edu/interface97.html ------------------------------ From: Rob Womersley Date: Tue, 22 Apr 1997 11:49:17 +1000 (EST) Subject: Position at New South Wales Centre Computational Scientist position at PVL, Sydney, Australia The New South Wales Centre for Parallel Computing (NSWCPC) and Sydney VisLab have established the Parallel computing and Visualization Laboratory (PVL) at the Australian Technology Park (ATP) in Sydney, Australia. A Computational Scientist position, with a focus on parallel computing, is available. The principal duties are research on projects associated with the PVL, presenting seminars to graduate students and staff, advising on efficient use of the facilities, development, maintenance and documentation of software. A PhD in Science or Engineering, a sound knowledge of Fortran, experience with UNIX and ability to communicate with users are essential. Experience with advanced computer architectures, Fortran 90, HPF, C programming, parallel scientific and engineering packages as well as scientific visualization and an understanding of communication networks and data archives are highly desirable. The position is for up to two years depending on the level of the appointment. See http://www.pvl.edu.au/jobs.html for more details. Enquiries may be directed to Dr. Rob Womersley, R.Womersley@unsw.edu.au. Dr. Rob Womersley E-mail: R.Womersley@unsw.edu.au School of Mathematics Phone: +61 - 2 - 9385-2998 University of New South Wales Fax: +61 - 2 - 9385-1071 Sydney, NSW, 2052 AUSTRALIA http://www.maths.unsw.edu.au/~rsw ------------------------------ From: Paul Tackley Date: Tue, 22 Apr 1997 12:50:24 -0700 Subject: Position at UCLA Postdoctorial Position/"Grand Challenge" Mantle and Core Dynamics The Institute of Geophysics and Planetary Physics at UCLA seeks a postdoc to work on a NASA HPCC "Grand Challenge" project for 3-D spherical modeling of convection in Earth's mantle and core. Project involves a multi-institutional (UCLA, Los Alamos National Lab., Johns Hopkins, UC Berkeley, U. Hawaii, Harvard) team of scientists, with the technical goal of acheiving performance (Gflop) milestones in 3-D spherical models of mantle and core dynamics on a Cray T3E, and the scientific goals of including 'coupling' between the core and mantle models, and incorporating plate tectonics into the mantle model. The successful applicant will work closely with Paul Tackley and Gerald Schubert at UCLA on implementing and researching strategies for modeling plate tectonics identified by the team, using multigrid codes. He/she may also work with spectral codes for geodynamo and mantle modeling. Applicant must be proficient in FORTRAN and knowledgable about numerical methods for fluid dynamics. Experience in massively parallel processing, and code performance tuning advantageous. Must have experience in mantle convection or other computational fluid dynamics. Position is 1 year renewable for another year, available immediately. Salary will be commensurate with qualifications and based on UC salary scales. To apply send resume and the names, addresses, phone numbers and email addresses of three referees to Paul Tackley, ESS, UCLA, 405 Hilgard Avenue, Los Angeles CA90095-1567, by May 15. For more information go to http://curie.eps.jhu.edu/info.html or contact Paul Tackley (ptackley@ess.ucla.edu, 310-206-9180) or Gerald Schubert (gschuber@ess.ucla.edu, 310-825-4577). UCLA is an equal opportunity/affirmative action employer. ------------------------------ From: Michael Hanke Date: Wed, 23 Apr 1997 12:13:31 +0200 (MET DST) Subject: Position at Humboldt-Unversity Berlin The Institute of Mathematics of the Humboldt-University of Berlin invites applications for a Research Position (Numerical Analysis) in a research project "Numerical methods for Schroedinger-Poisson systems for the modelling of semiconductors" funded by the German Research Foundation. The successful apllicant will take part in the development and implementation of efficient algorithms for the Schroedinger-Poisson system. We look for a numerical analyst with high knowledge and experience in the field of numerical methods for partial differential equations as well as eigenvalue problems. Skills in designing scientific software are necessary. Basic knowledge of semiconductor physics could be useful. The position is available for 2 years and can be filled immediately. The salary and the social benefits are in accordance with the German salary scale BAT-O IIa/2. For further information please contact Michael Hanke (hanke@mathematik.hu-berlin.de, na.mhanke@na-net.ornl.gov). Application should be sent to PD Dr. Michael Hanke Humboldt-University of Berlin Institute of Mathematics D-10099 Berlin Germany ------------------------------ From: Nick Higham Date: Wed, 23 Apr 1997 11:44:16 +0100 Subject: Studentship at University of Manchester EPSRC CASE Ph.D. Studentship for project Numerical Solution of the Symmetric Definite Generalized Eigenproblem in the Department of Mathematics, University of Manchester. This three-year CASE studentship is funded by the Mathematics Committee of the Engineering and Physical Sciences Research Council (EPSRC) with the Numerical Algorithms Group (NAG) Ltd., Oxford as the cooperating body. The studentship leads to the award of a Ph.D. degree and is available from September 1997. The student will receive the usual EPSRC Research Studentship grant plus an additional \pounds 2,850 per annum. The project will be supervised by Professor N. J. Higham in cooperation with Professor S. J. Hammarling of NAG. It involves the development, analysis and implementation of a numerical method for solving an important class of algebraic eigenvalue problems arising in many engineering applications. Applicants should have or expect a good (I or II(i)) honours degree in Mathematics or in a programme which includes substantial work in mathematics. CASE students are normally fully funded only for UK residents, but EC students may be funded on a "fees only" basis. For further information and applications please contact Professor Nicholas J. Higham Department of Mathematics University of Manchester Manchester M13 9PL Tel. 0161 275 5822 email: higham@ma.man.ac.uk http://www.ma.man.ac.uk/~higham/ ------------------------------ From: Eric Grosse Date: Thu, 24 Apr 1997 09:51:07 +0400 Subject: Postdoctoral Position at Bell Laboratories POSTDOCTORAL POSITION: COMPUTER MODELING OF SURFACE EVOLUTION Bell Laboratories, Murray Hill, NJ Applications are invited for a postdoctoral position at Bell Laboratories in Murray Hill, New Jersey. The work will involve numerical modeling of thin film deposition using both continuum and discrete, atomic-level, Monte Carlo techniques. The inclusion of stochastic aspects of the atomistic deposition process into a general continuum model is a major goal of this work. Candidates should have strong backgrounds in numerical analysis and computer programming, with experience using C or C++. The position will be supervised by the applied mathematics research and the silicon processing research areas. The work will involve frequent interactions with materials scientists working on various aspects of computer modeling and experimentalists testing the predictions of the models in laboratory apparatus and in actual fabrication lines for silicon devices. Effective communication skills and a broad range of interests are essential. The position is jointly funded by the NSF Divisions of Mathematical Sciences and Materials Research, and by the DARPA Defense Sciences Office. A recent Ph.D. in applied mathematics is required. The position is open immediately, and is funded for two years. The position requires US citizenship or permanent resident status. The position will be closed as soon as a suitable candidate is found. Curriculum vita should be sent first by e-mail, and later three letters of reference. George Gilmer 1E-434 Bell Laboratories, Lucent Technologies Murray Hill, NJ 07974 Phone: (908) 582 5547. Email: ghg@lucent.com Frieder Baumann Bell Laboratories, Lucent Technologies Room 1E-234 Murray Hill, NJ 07974 Phone 908-949-1362. Email: frieder@lucent.com Lucent Technologies is an affirmative action/equal opportunity employer and encourages applications from minorities and women. ------------------------------ From: Fran Moshiri Date: Thu, 24 Apr 1997 09:41:14 -0500 Subject: Research Faculty Position at Rice University RICE UNIVERSITY Department of Computational and Applied Mathematics The Department of Computational and Applied Mathematics invites applications for a research faculty position beginning in the fall semester of 1997. The position will be at the rank of faculty fellow, senior faculty fellow, or distinguished faculty fellow depending on experience and qualifications. Applicants should hold a Ph.D. degree or the equivalent in computational and applied mathematics or a related discipline. A candidate for this position must have established an exceptional research record in the areas of programming tools, automatic differentiation, and problem solving environments. Rice University is a private research university with a long tradition of excellence in undergraduate science and engineering education. The Computational & Applied Mathematics Department hosts research programs in Linear and Integer Programming, Numerical Optimization, Numerical Linear Algebra, Parallel Computing, Optimal Design and Inverse Problems for Partial Differential Equations. Applicants should send a letter of application; a brief statement of research accomplishments, interests, and goals; a curriculum vitae and selected reprints; and have three letters of recommendations sent to: Chair, Staffing Committee Department of Computational & Applied Mathematics-MS 134 Rice University 6100 Main Street Houston, Texas 77005-1892 APPLICATION DEADLINE: May 31, 1997 AN AFFIRMATIVE ACTION, EQUAL OPPORTUNITY EMPLOYER Franka Moshiri |Phone (713) 737-5627 Department Administrator |Dept Fax (713) 285-5318 CAAM Department-MS 134 |Dept Phone (713) 527-4805 Rice University | 6100 Main Street | Houston, Texas 77005-1892 |Office: 1078 Duncan Hall ------------------------------ From: Karl Kunisch Date: Thu, 24 Apr 1997 17:37:37 +0200 Subject: Position at University of Graz INSTITUTE OF MATHEMATICS KARL FRANZENS UNIVERSITAET GRAZ, AUSTRIA At the Institute of Mathematics at the University of Graz there is an opening for a position as "University Assistent". Requirements: University degree in mathematics Qualifications: Solid knowledge of numerical analysis, partial differential equations, preferably knowledge of fluid mechanics, optimization, UNIX. The first contract is for two years. It is renewable. Deadline for application: May 7, 1997 Reference number: 23/109/99 Address: Dean of the faculty of science, Universitaet Graz Universitaetsplatz A-8010 Graz, Austria The German version of the official text follows: 1 Planstelle fuer eine Universitaetsassistentin oder einen Universitaetsassistenten am Institut fuer Mathematik ab sofort. Aufnahmebedingungen: Abgeschlossenes Hochschulstudium aus Mathematik beliebiger Studienrichtung oder Studienzweiges. Erwuenschte Kenntnisse bzw. Qualifikationen: Fundierte Ausbildung in Numerischer Mathematik und partiellen Differentialgleichungen, vorzugsweise Gleichungen der Fluidmechanik; Kenntnisse aus Optimierungstheorie (stetige), UNIX-Systemerfahrung. Ende der Bewerbungsfrist: 7. Mai 1997 (Kennzahl: 23/109/99). For further information please feel free to contact: Prof. Karl Kunisch | Email karl.kunisch@kfunigraz.ac.at Institut fuer Mathematik | Phone 43-(0)316 380-5162 University of Graz | Fax 43-(0)316 380-9815 Heinrichstrasse 36 | A-8010 Graz, Austria ------------------------------ From: Ivan Graham Date: Fri, 25 Apr 1997 17:25:27 +0100 Subject: Studentship at University of Bath Postgraduate Studentship - University of Bath School of Mathematical Sciences Applications are invited for a three year postgraduate studentship funded under the UK Engineering and Physical Sciences (EPSRC) Cooperative Awards in Science and Engineering (CASE) Scheme. The cooperating body is AEA Technology, Harwell, Oxford, UK and funds are available to support a PhD student working on the topic: Parallel Algorithms in Groundwater Flow This project is concerned with the parallel solution of linear systems arising from finite element discretisations of two- and three-dimensional models for groundwater flow coupled with the transport of salinity. AEA Technology has written and markets the code NAMMU (``Numerical Assessment Method for Migration Underground'') which is used commercially on a range of pollution control applications. Mathematically the project concerns the development and implementation of parallel iterative methods for mixed finite element discretisations of partial differential equations describing groundwater flow, and in particular for the coupled system which arises when nonlinear salinity effects are included. Compuationally it is envisaged that the student will implement these methods using the approach of the DOUG (Domain Decomposition on Unstructured Grids) package recently developed at Bath. Details of this package, which uses the message passing system MPI, are available at http://www.maths.bath.ac.uk/~mjh/doug. At present this package solves standard finite element discretisations of PDEs in 2D and 3D. One of the chief initial computational tasks will be to extend it to mixed finite elements. Computing for the project will be done at the University of Bath and at AEA Technology, to which the student will make regular scientific visits. The remuneration provided by this studentship will be 5,295 pounds from EPSRC plus 2,355 pounds from the cooperating body and the student will also have the opportunity to undertake some paid part-time tutoring duties at the University of Bath. Applicants for this studentship should have or expect to obtain the equivalent of a 2(i) Honours Degree or better in a numerate discipline, and should have some experience in numerical analysis, differential equations and numerical computation. Applications from suitably qualified students with mathematics, engineering or computer science backgrounds are all welcome. Interested applicants should please contact Dr I.G. Graham, School of Mathematical Sciences, University of Bath, Bath BA2 7AY, United Kingdom. Phone: +44, 1225 826343, or +44, 1225 826989. Fax: +44, 1225 826492. Email: I.G.Graham@bath.ac.uk, http://www.maths.bath.ac.uk/~igg ------------------------------ From: Ivan Graham Date: Fri, 25 Apr 1997 17:29:58 +0100 Subject: Postdoctoral Position at University of Bath Postdoctoral Research Assistant - University of Bath School of Mathematical Sciences Applications are sought for a United Kingdom EPSRC funded one-year position to work on domain decomposition techniques for the iterative solution of elliptic PDEs on unstructured grids. The project is to extend the work of a previous project which produced the ``DOUG'' package for parallel iterative solution of elliptic problems, details of which can be found at http://www.maths.bath.ac.uk/~mjh/doug. The DOUG package solves PDEs by iterative methods with domain decomposition preconditioning and is implemented in parallel using MPI. Very good speed-ups have been obtained on dedicated parallel machines such as the SP/2, as well as on ethernet clusters. At present full additive Schwarz preconditioners are implemented and tested in 2D only. The main aim of the present project is to extend the full preconditioning technique to 3D problems, and in particular to implement and test an appropriate unstructured coarse grid generation procedure in that case. For this work applicants should have an interest and experience in large scale scientific programming, preferably in Fortran. Experience in parallel programming using message passing techniques would be a distinct advantage. Possible extensions of the main aim of the project include the parallel solution of mixed finite element discretisations of groundwater flow problems arising from industrial applications and/or extensions of a recent theory describing the performance of Schwarz methods in the presence of highly discontinuous coefficients in irregular geometries. Related theoretical work arising from the DOUG project can be viewed at Reports 9608 and 9619 on http://www.maths.bath.ac.uk/Maths_Group/preprints.html The position is available on or as soon as possble after September 1st 1997 for one year. The appointment will be at the Grade 1A level and the salary will be around 16.6K pounds, subject to confirmation and depending on age and experience. Further details of the project may be obtained from the supervisor, Dr Ivan Graham, School of Mathematical Sciences , University of Bath (email: igg@maths.bath.ac.uk, telephone +44 (225) 826989, fax +44 (225)826492, http://www.maths.bath.ac.uk/~igg). ------------------------------ From: Thomas Liebling Date: Sat, 26 Apr 1997 14:31:19 +0000 Subject: Positions at EPFL, Lausanne, Switzerland The Swiss Federal Institute of Technology Lausanne (EPFL) has four openings for a MAITRE D'ENSEIGNEMENT ET DE RECHERCHE (MER) (SENIOR LECTURER): 1.OPERATIONS RESEARCH at the Mathematics Department The activities of the new collaborator will involve all aspects of Operations Research (modelling, simulation and optimization). He/she will be assigned to one of the Chairs of the Operations Research Group of the Mathematics Department. 2.NANOPHYSICS at the Physics Department The new collaborator will lead a research team on scanning tunnelling microscopy in the Institute of Experimental Physics. The research activity of the candidate will primarily focus on the development of scanning tunnelling spectroscopy into a chemical sensitive probe with high spatial and temporal resolution. The candidate should have developed a high level of research in experimental condensed matter physics, chemical physics or a related field. An outstanding research record and a commitment to excellence in teaching are advantages. Extensive research experience in the field of surface science and scanning tunnelling microscopy is required. Experience with nanostructured surfaces is desirable. 3.CHEMICAL REACTION ENGINEERING at the Chemistry Department The new collaborator is expected to develop an original research program at the leading international level in chemical reaction engineering. The desired area of research concerns heterogeneous catalysis in connection with the design of microreactor systems. He/she must have proof of his/her ability to develop and lead an outstanding research program. Industrial experience is an advantage. 4.PHYSICAL CHEMISTRY at the Chemistry Department The new collaborator is expected to develop an original research program at a high level. He/she will propose and develop activities in the domainof "ultrafast spectroscopy and molecular reaction dynamics". For the four positions: Activities will be centered at the concernedDepartments, but they will also involve other units of EPFL, as well as further Swiss and international academic institutions and private companies. Askill for teaching graduates and undergraduates and for conducting original and high level research is essential. The new collaborators are also expected to supervise and guide students on their semester projects and on their diploma and PhD theses. They should possess a marked taste for project leadership. Applications are encouraged from people who fulfill the requirements of the Swiss program for ensuring the continuity of competent university faculty. Deadline for applications: August 15, 997. Starting date: as mutually convenient. Applications from women are particularly welcome. For further information, please contact by writing: Presidence de l'Ecole polytechnique federale de Lausanne CE-Ecublens, CH 1015 Lausanne, Switzerland ------------------------------ From: IEEE Computer Date: Mon, 21 Apr 97 16:35:13 est Subject: Contemts, IEEE Computational Science & Engineering Contents of IEEE COMPUTATIONAL SCIENCE & ENGINEERING magazine, Vol. 4, No. 1, Jan.-Mar. 1997. http://computer.org/cse Editor-in-Chief: George Cybenko, Dartmouth College, george.cybenko@dartmouth.edu Associate EIC: Francis Sullivan, IDA Ctr. for Computing Sciences THEME SECTION ARTICLES--Microelectromechanical Systems (MEMS) and Computation Distributed MEMS: New Challenges for Computation Andrew A. Berlin and Kaigham J. Gabriel 12 Computational Methods for Design and Control of MEMS Micromanipulator Arrays Karl F. Borhinger, Bruce R. Donald, Noel C. MacDonald, Gregory T.A. Kovacs, and John W. Suh 17 Simulating the Behavior of MEMS Devices: Computational Methods and Needs Stephen D. Senturia, Narayan Aluru, and Jacob White 30 Challenges in Commercializing MEMS Eric Peeters 44 Developing Infrastructure to Mass-Produce MEMS Karen W. Markus 49 FEATURE ARTICLES Emulating Shared Memory to Simplify Distributed-Memory Programming Jerry A. Clarke 55 The Visual Interactive Desktop Laboratory Upul Obeysekare, Fernando F. Grinstein, and Gopal Patnaik 63 Distributed Edge Detection: Issues and Implementations Xiaodong Zhang, Sandra G. Dykes, and Hong Deng 72 DEPARTMENTS Associate Editor-in-Chief 1 Will the Circle Be Unbroken? (A Meditation on the Passing of Seymour Cray) --Francis Sullivan Editorial Board Changes -- Letters CSE Education 7 SCCM Advice: Student-Run Math Consulting at Stanford --Melissa Aczon, Martin Gander, Margot Gerritsen, Tony Shardlow, and Ronnie Sircar About the Articles 10 Book News & Reviews 83 Bookshelf --E. Gallopoulos _High Performance Computing: Challenges for Future Systems_, by David J. Kuck; reviewed by John R. Gurd Technology News & Reviews 87 Fortran 90 news, free software, symbolic computation packages, Matlab 5, Web products from Visual Numerics --R. Bramley 1996 Referees 91 Computing Prescriptions 92 Dynamic Programming: A Discrete Calculus of Variations --George Cybenko Conferences & Workshops 98 Int'l Conf. on Parallel Computing [Minnesota] --Donald G. Truhlar High-Performance Computer Applications in the Behavioral Sciences --Lynne K. Edwards Frontiers of Massively Parallel Computation '96 --Dan Pryor Call for Articles 102 IEEE CS&E Theme Section on Problem-Solving Environments for Scientific Computing (submissions due May 15, 1997) IEEE CS&E Theme Section on Geometric Hashing (submissions due May 15, 1997) Interfaces 104 What Shapes Computing's Future? More Than Hardware --Norris Parker Smith ------------------------------ From: Baltzer Science Date: Fri, 25 Apr 1997 08:57:10 +0200 Subject: Contents, Computational Geosciences Computational Geosciences 1 (1997) 1 Laurent Ann'e, Quang Huy Tran and William W. Symes Dispersion and cost analysis of some finite difference schemes in one-parameter acoustic wave modeling 1-33 J.M. Guevara-Jordan and J. Glimm A mixed finite element method for Hele-Shaw cell equations 35-58 Andro Mikelic and Laetitia Paoli On the derivation of the Buckley--Leverett model from the two fluid Navier--Stokes equations in a thin domain 59-83 W.A. Mulder Experiments with Higdon's absorbing boundary conditions for a number of wave equations 85-108 ------------------------------ From: Oleg Burdakov Date: Fri, 25 Apr 1997 19:07:33 +0200 (MET DST) Subject: Contents. Optimization Methods and Software Table of Contents Optimization Methods and Software (OMS) Volume 7, Numbers 3=4 (May, 1997) M. Dobmann and K. Schittkowski Constrained parameter estimatuion in partial differential equations 165=210 M. Heinkenschloss The numerical solution of a control problem governed by a phase field model 211=263 M. Anitescu, G. Lesaja and F.A. Potra Equivalence Between Different Formulations of the Linear Complementarity Problem 265=290 B. He, E. de Klerk, C. Roos and T. Terlaky Method of approximate centers for semi=definite programming 291=309 I. Konnov A Combined Method for Smooth Equilibrium Problems with Constraints 311=324 ------------------------------ End of NA Digest ************************** -------