Roberto, Margherita and Beatrice


Personal Info




On the Detection of Implicit and Redundant Numeric Constraints in CLP Programs

Roberto Bagnara
Dipartimento di Informatica
Università di Pisa
Corso Italia 40
I-56125 Pisa


We address the problem of compile-time detection of implicit and redundant numeric constraints in CLP programs. We discuss how this kind of constraints have several important applications in the general field of semantics based program manipulation and, specifically, optimized compilation. We attack the problem in an original and effective way. The basic ideas underlying our approach are: (1) the use of approximate and efficient constraint inference techniques originally developed in the field of artificial intelligence. These techniques allow great flexibility in dealing with the complexity/precision tradeoff of the analysis. And (2) the integration of these techniques within an appropriate framework for the definition of non-standard semantics of CLP. We show that one notable advantage of this combined approach is that it allows to close the often existing gap between the formalization of data-flow analysis in terms of abstract interpretation and the possibility of efficient implementations. Some preliminary results (both in terms of complexity and precision) obtained with our prototype implementation are presented.

Available: PDF, 300 DPI, 600 DPI, and 1200 DPI PostScript, DVI, BibTeX entry.

[Page last updated on December 10, 1999, 11:30:49.]

© Roberto Bagnara

Home | Personal | Papers | Teaching | Links