http://interval.louisiana.edu/conferences/Validated_computing_2002/html_notice.html

Validated Computing 2002

A SIAM Workshop, May 23-25, 2002

immediately following the

Seventh SIAM Conference on Optimization

Click here to register

/ General information / Call for papers / Deadlines / Pre-invited speakers / Informal week at the Fields Institute /  Description and rationale / Program committee / Schedule with links to abstracts / Moore Prize / Registration and Accomodation / Proceedings submissions / Abstracts alphabetically by author /

General information

Validated Computing 2002 will deal with all aspects on validated computing, although we hope for a special emphasis on applications and on tools to support validated computations. There is also a special opportunity to foster interaction between validated computing experts and the optimization community at large.  There will be a special session honoring the contributions of Ramon Moore.

For more detailed thoughts on these goals and on validated computing, see the description and rationale.

Call for papers

We invite submission of papers dealing with validated computing. Areas of interest include, but are not limited to, interval arithmetic and analysis, use of mathematical theory to assure reliable scientific computation, fuzzy interval analysis, and fuzzy logic.  We are especially interested in applications of these techniques and in tools that support the techniques.

For consideration, submit an extended abstract of 2-3 pages to R. Baker Kearfott  at rbk@louisiana.edu by February 15.  Submissions must be in Latex using only the standard article style.  Use psfig.sty if you include figures.

The Program Committee will review the submissions.  Papers will be presented as 30 minute talks.  A few talks may be selected for special sessions.
 

Deadlines

 
February 15 Extended 2-3 page abstracts for submissions.
March 1 Notification of accepted talks
March 15 (tentative)  End of early registration
March 15 Revisions to extended abstracts due
April 1 (tentative)  Hotel reservations
May 23--25  Workshop itself

A few of the people already committed to speak:

Informal week at the Fields Institute

Participants in the workshop are invited to participate in a week of informal discussions at the Fields Institute at the University of Toronto, immediately following the workshop. This week is part of the Fields Institute's special thematic year on computational challenges in science and engineering.   We  hope to foster interaction and cross-fertilization between validated optimization experts and experts in "standard"  optimization techniques. This will be primarily an informal, colleague-to-colleague  meeting to complement the SIAM Conference on Optimization and Validated Computing 2002.  Depending on numbers, the Fields Institute can provide office space and meeting areas for this activity. For general information on the Fields Institute and the thematic year, see  http://www.fields.utoronto.ca/programs/scientific/01-02/numerical/).  For more information on our informal meeting, see
http://www.fields.utoronto.ca/programs/scientific/01-02/numerical/optimization/.

Persons interested in this week at the Fields Institute should contact R. Baker Kearfott (rbk@louisiana.edu) or
or Tibor Csendes (csendes@inf.u-szeged.hu).

Description and rationale

Reliable computing techniques are essential if the validity of a solution is to be guaranteed. There is no feasible alternative. Modern societies rely more and more on computer systems. Usually, our systems appear to work successfully, but there are sometimes serious, and often minor, errors. Ever increasing reliance on computer systems brings ever increasing need for reliability.

Validated computing is one essential technology to achieve increased software reliability. Validated computing uses controlled rounding of computer arithmetic to guarantee that hypotheses of suitable mathematical theorems are (or are not) satisfied. Mathematical rigor in the computer arithmetic, in algorithm design, and in program execution allow us to guarantee that the stated problem has (or does not have) a solution in an enclosing interval we compute. If the enclosure is narrow, we are certain that we know the answer reliably and accurately. If the enclosing interval is wide, we have a clear warning that our uncertainty is large, and a closer study is demanded.

Intervals capture uncertainly in modeling and problem formulation, in model parameter estimation, in algorithm truncation, in operation round off, and in model interpretation.

The techniques of validated computing have proven their merits in many scientific and engineering applications. They help answer questions from, "How much irrigation water does a desert golf course return effectively unused to its bordering stream?" to "Will a near earth asteroid hit the earth, possibly ending life as we know it?".

The techniques of validated computing rest on solid and interesting theoretical studies in mathematics and computer science. Contributions from fields including real, complex and functional analysis, semigroups, probability, statistics, fuzzy interval analysis, fuzzy logic, automatic differentiation, computer hardware, operating systems, compiler construction, parallel processing, and software engineering are all essential.

The major emphasis of the program is on applications. We will hear from many people who have used tools from validated computing to attack, and often solve, significant practical problems. Successful applications have included medical diagnosis and treatment, financial simulation, mechanical design, oil reservoir simulation, aeronautics, high energy particle accelerators, environmental engineering, chemical process simulation and control, computer graphics for motion picture special effects, astrophysics, and many more.

Not all applications are as yet successful.  We will also hear from people with challenging applications to which validated techniques have not yet been successfully applied.  Hopefully, by encouraging experts in such applications to lay out their problems, we will foster long-term collaborations leading to significant advances in those fields.

The workshop follows the SIAM Optimization meeting because global optimization is a major concern of both the optimization and the validated computing communities.  By holding the meetings consecutively, we encourage validated computing researchers to become more involved in the wider optimization community, and we encourage people more interested in standard techniques of optimization to participate in interval discussions.

We will have one special session and a conference banquet to honor Ray Moore.  His 1966 book defined the field, he pioneered many applications, and he continues to contribute insights and papers. Most of the ideas in our interval algorithms of today directly trace their ancestry to Ray's 1966 and 1979 (from SIAM) books.

In parallel with the traditional scientific program following SIAM's usual pattern of highlighted and contributed papers, we are considering half-day detailed workshops.  Tentative topics include:
 

  1. Jiri Rohn on complexity.  This would follow up on his talk at SCAN 2000 in Karlsruhe "Finite Characterization of Some Linear Problems with Inexact Data.
  2. Tutorial on validated techniques, interval arithmetic, and related tools.  We would start at the beginning by defining directed rounding, and progress to a "Numerical Recipes" level view of several widely used algorithms, e.g., linear systems, interval Newton, global optimization, ordinary and partial differential equations.
  3. Hands-on tools and demonstrations.

Program committee

Moore Prize

The biennial Moore Prize for Applications of Interval Analysis will be awarded for the first time at Validated Computing 2002. For details, including how to place a work into consideration for this prize, see http://interval.louisiana.edu/Moore_prize.html.

Registration and accomodation

For full hotel information, as well as an on-line form for registration, go to

http://www.siam.org/meetings/vc02/index.htm

Proceedings submissions

Full versions of papers presented at Validated Computing 2002 will be published in one or more special issues of the journal Reliable Computing after the conference.  These papers will go through the refereeing and editorial process for Reliable Computing.  To be considered for one of these special issues, papers should be sent on or before July 1, 2002 to
R. Baker Kearfott
Editor, Validated Computing 2002 Proceedings
Department of Mathematics
University of Louisiana at Lafayette
Box 4-1010
Lafayette
LA 70504-1010
USA