PRINCE (EP 5246)

Prolog Integrated with Constraints and Environment for Industrial and Financial Applications

This page contains some information regarding the project, with some additional detail on the contributions of the academic partners (in the areas of program analysis and transformation). These partners can be contacted for more information on these topics. For more information on the project in general please contact the project coordinator or the other partners.

COORDINATOR

PrologIA (France)

CONTACT POINT

Guy Alain Narboni (PrologIA)
Tel. +33-91 268636
Fax +33-91 41 96 37

PARTNERS and ASSOCIATED PARTNERS

R. Bosch GmbH (Germany): Reinhard Weber
Tel. +49 711-811 6384
Fax +49 711-811 6800
CEFI, Faculté des Sciences Economiques (France): Christophe Bisire
Tel. +33 4227 4266
Fax +33 4238 9585
Banque La Hénin (France): Gilbert Osio
Tel. +33 1-4451 2020
Fax +33 1-4451 2839
Katholieke Universiteit Leuven, Dept. of Computer Science (Belgium): Maurice Bruynooghe
Tel. +32 1-620 1015
Fax +32 1-620 5308
University of Bristol, Dept. of Computer Science (United Kingdom): John Gallagher
Tel. +44 272-303 307
Fax +44 272-251 154
Universidad Politécnica de Madrid, Facultad de Informática (Spain): Manuel Hermenegildo
Tel. +34 1-336 7435
Fax +34 1-3524819
FAW (Germany)DASA AG (ex MBB) (Germany): Reinhard Skuppin
Tel. +49 731-501 413
Fax +49 731-591 999
DASA AG (ex MBB) (Germany): Thomas Buckle
Tel. +49 89-607 27892
Fax +49 89-607 23583

Start Date: September 1990

End Date: January 1995

Duration: 53 months

Status: completed

Abstract

The Prince project is concerned with constraint logic programming. It is developing a new constraint logic programming (CLP) language heavily based on PROLOG III technology.

RATIONALE

In order to apply the technology to large-scale applications, performance, robustness and expressiveness must be improved. This is the aim of Prince. Its general objectives are: to bring a 'PROLOG III'-like system to a high level of industrial applicability and usefulness, and to validate and augment the industrial significance of such systems by applying them effectively to real-life applications.

GENERAL DESCRIPTION AND APPROACH

This project is a continuation of the Esprit I project P1106. Its aim was the further development of Prolog and its validation by KBS in technical areas. It demonstrated the potential of the Constraint Logic Programming (CLP) approach through the successful development of Prolog III. To be applicable to large-scale applications, however, such a system must still gain in performance, robustness, and expressiveness. The general objectives of this project are threefold:

to bring a 'Prolog III'-like system to a higher level of industrial applicability and usefulness by means of global compilation, improvements to constraint-solving algorithms, more specific constraint domains, and better programming environments.
to validate and augment the industrial significance of such systems by applying them effectively to real-life applications provided and developed by IT users who are partners in this project. The application work will concern the areas of finance, industrial systems engineering, and production planning and scheduling.
to integrate the experience of a major Prolog manufacturer, PrologIA, into a new joint system which would then have the means to become a significant industrial force. PrologIA has developed the Prolog II+ compiler and Prolog III, a major example of a constraint logic programming language.

APPLICATION DESCRIPTIONS

Application in the financial domain. The development of a forecasting tool which evaluates bank profits on the bases of, on the one hand, certain techniques currently used for the projection of balance sheets and, on the other hand, the analysis of strategic planning taking into account hypotheses concerning the modifications of the environment as they might arise in years to come.
Industrial systems engineering. The chosen application for this project demands substantial extensions to the model-based diagnosis methodology developed in a previous project. Two technical application fields will serve as testbeds for the new language:
- System analysis with respect to reliability and safety of automotive systems;
- Fault detection and isolation in a satellite sub-system. Scheduling application.
This last task will concentrate on the two main issues raised in the project description:
- the application of CLP to algorithms for Prince Prolog System (PPS) problems (e.g. resource based approaches);
- the improvement of the scheduling algorithm to make it applicable to as great a variety of industrial scheduling problems as possible (this includes topics such as scheduling with intermediate products, aggregation of orders, partial execution of orders, replanning, and possibly others).

Work to date, problems encountered and remaining activities

The kernel of the Prince Prolog system (Prolog IV) is now finished. It includes a compiler, as well as the following solvers:

Equations and disequations over rational trees.
Linear numerical constraints using Gaussian elimination and Simplex algorithms.
Numerical constraints using intervals.

A set of analysis and program transformation tools has also been developed. These tools allow performing several optimizations on source programs prior to compilation with the objective of increasing the performance of the resulting compiled code. In addition, the activities of the project have produced numerous theoretical results in the areas of language design and implementation, constraint solving algorithms, global analysis, and program transformation for constraint logic programming languages. Current activities include:

Conformity with the ISO Prolog draft standard.
Completion of the development environment.
Solver optimization.

The first version of the product will be marketed by March 1995, over a limited range of platforms. More ports will be performed by June 1995.

ROLES OF PARTNERS

Compilation of the standard Prolog part of Prolog III: PrologIA.
Compilation of constraints and first integration of the system: PrologIA.
Pre-compilation and global analysis: KUL, Bristol, and Madrid universities.
Development of specific constraint domains for industrial applications and improvement of existing algorithms: PrologIA.
Development environment and software engineering: PrologIA.
Validation through industrial applications: BOSCH, Banque la Henin, FAW, CEFI, DASA (ex MBB).
Integration into industrial system: PrologIA.
Management and information dissemination: PrologIA, KUL, BOSCH, FAW.

EXPLOITATION

PrologIA: Co-ordinating contractor. Distribution of a user-friendly, efficient, compiled version of the new CLP Language which will satisfy the needs of industrial users, computer service companies involved in medium term applied research and universities. An industrial version of the product that includes the specifications required by industry will have a much wider potential market. The production of applications by means of this tool.
Bosch: Partner, industrial interest. Getting further experience on CLP and its industrial potential. Efficient systems engineering tools for complex technical systems and diagnosis especially in vehicles and autonomous mobile systems. Shorter development cycles. Applying linear constraint programming in different fields of system engineering: FMEA, fault detection and isolation.
La Henin Banque: Associated Partner, industrial repercussion. The proposed software is of great industrial interest as banks have a growing need for general evaluation systems in connection with the assessment of the growth of interest rates and in general risk factors.
DASA (ex: MBB): Associated Partner, industrial interest. MBB is interested in investigating the applicability of CLP-base methods to complex dynamic satellite control sub-systems, especially for diagnostic purposes. The experience with extended CLP can support and improve the development of future ground test facilities.
CEFI, Katholieke Universiteit Leuven, University of Bristol and University of Madrid: Associated Partners. The academic partners intend to exploit the results of the project in a non-commercial way. The results of the non-confidential work-packages will be published, and used in their doctoral curriculae. They will use the results as a basis for continued research on abstract interpretation, program transformation and constraint solving in Logic Programming.

Other Comments

BIM (Belgium) was initially a partner of the Prince Project, involved in the compilation of the standard Prolog. However it left the consortium before the end of the project.

<webmaster@clip.dia.fi.upm.es>