Section: New Results

Execution models for Constraint Programming and Semantics Equivalence

Participants : François Fages, Thierry Martinez, Sylvain Soliman.

Logic programming and constraint programming are two declarative programming paradigms which rely on the identification of programs to theories, and programming to modeling. Execution models result from the operational interpretation of logical provability in logic programming, and of constraint propagation in constraint programming. However, the control of execution is crucial for the practicability of these schemes and extra-logical traits are thus added in those programming systems, with the classical slogans "logic program = logical theory + control", "constraint program = constraint model + search".

In his thesis [4] , Thierry Martinez investigates execution models in which control and search can be shifted into the logic or the constraint model, while preserving the semantics. The three parts of the thesis correspond to the three semantics equivalence that are showed: the first between two committed-choice forward-chaining logic languages, the second between constraint logic programs and constraint models, and the third between guard semantics in angelic settings. Each of these equivalence is constructive in the sense that there exists an encoding that enables the compilation from one of the paradigm to the other.