the sac newsletter
Number 1, November 1996.

Ongoing European Projects in Computer Algebra


The FRISCO Project

Mike Dewar

NAG Ltd
Wilkinson House, Jordan Hill Rd
Oxford, OX5 1DP, United Kingdom
Email: miked@nag.co.uk


FRISCO: A Framework for Integrating Symbolic/Numeric Computation is an Esprit Long Term Research project which began in March 1996. The project lasts three years and is dedicated to developing state-of-the-art algorithms for solving systems of polynomial equations, and a framework for delivering implementations of those algorithms in an industrial environment. The FRISCO consortium consists of the Numerical Algorithms Group Ltd (NAG) as co-ordinating partner, with INRIA Sophia-Antipolis, IRMAR at the University of Rennes, the University of Cantabria and the University of Pisa as partners. This consortium brings together some of the main members of the PoSSo consortium, who under a previous Esprit project developed amongst other things a C++ library for solving polynomial systems, with some of the people directly involved in developing Aldor, the new compiler for the AXIOM computer algebra system.

While polynomial problems arise in many areas of science and engineering - robotics, telecommunications, scheduling to name but three - until now there has been very little use made of computer algebra in these areas. This is partly because the interactive environments currently available for computer algebra are not capable of solving very large problems quickly and efficiently, and partly because they cannot be interfaced effectively to other scientific applications written in more conventional languages like C and Fortran. The FRISCO project aims to address these issues by, on the one hand, producing very efficient implementations of algorithms tailored for problems with specific characteristics and, on the other hand, developing a framework for these algorithms which will allow them to be embedded into applications easily and effectively.

There will be three technical strands to the project:

runtime
devoted to developing a common model for low-level runtime facilities for computer algebra;
interoperability
concerned with issues to do with mixing algebraic code written in Aldor with languages like C, C++ and Fortran; and
algorithms
concerned with conducting research into, and producing implementations of, algorithms to meet the perceived needs of industrial users.
In addition there will be a strand devoted to software integration and another for developing demonstration applications. At the end of the project it is hoped that a comprehensive framework for using computer algebra in an industrial framework will have been developed, and populated with tools for solving polynomial problems.

The project aims to be reactive to what it perceives to be the needs of European Industry, and held a workshop in Barcelona from 23rd to 25th of October 1996 which brought together industrial researchers with specialists in computer algebra. It is intended to maintain such contacts throughout the lifetime of the project.

Further information on FRISCO can be located at http://www.nag.co.uk/projects/FRISCO.html or by email from the project co-ordinator miked@nag.co.uk.


contentsContents