07/12/2017

Intervenant(s) : Pierre TALBOT (Repmus, Ircam)

Constraint solvers are complex pieces of software for solving and optimizing a set of constraints. Because these problems are NP-complete, the user must often create customized search strategies when tackling large problem's instances. However, it is notoriously complex for non-expert users to program and integrate search strategies into the solver. To solve this problem, we propose spacetime programming, a paradigm based on complete lattices and synchronous programming. It makes use of the time dimension of synchronous languages to program and compose search strategies.

In this talk, we will start with an overview of spacetime and an example of a search strategy. In a second part, we will focus on the semantics rules supporting the lattice-based view of variables. This comes as an extension of Esterel where variables can be defined over complete lattices instead of boolean signals.

Emmanuel.Chailloux (at) nulllip6.fr