Pascal Van Hentenryck, Brown University
Constraint Programming at work
Constraint programming is a success story. In the late 1980s, it moved from research laboratories to industrial applications in a few years and is now in daily use to solve complex optimization throughout the world. At the same time, constraint programming has continued to evolve, addressing new needs and opportunities. This talk reviews the essence of constraint programming and some recent progress, its hybridization with other optimization approaches, the quest for more autonomous search, and its applications in a variety of nontraditional areas.
Robert Nieuwenhuis, Technical University of Catalonia
The Barcelogic approach to search: fast and robust but expressive
During the last years we have developed theory and tools for SAT and SAT Modulo Theories (SMT), with emphasis on verification tasks. In this talk we describe our current work on applications to hard combinatorial problems in general.
We first give an overview of our DPLL(T) approach to SMT [JACM Nov'06] and its simplementation in the Barcelogic SMT tool. Then we discuss advantages inherited from SAT techniques (efficiency, robustness and lack of need for tuning) and see how it can also incorporate features of CP techniques such as rich modeling languages and the most efficient special-purpose algorithms for handling, e.g., linear arithmetic or global constraints. We finish with some rather surprising experimental results.