Expressiveness and Utility of Generic Representation Schemes

We describe how problems can be solved by “compiling” them to a general-purpose representation scheme such as SAT or PseudoBoolean (PB). We discuss why PB is a so much more expressive and better language than SAT for problems that involve any arithmetic reasoning. We describe PARIS, a program that solves decision problems expressed in PB, that is based on an adaptation of the best SAT solvers. We also describe OPARIS, an extension of PARIS to solve optimization as well as decision problems. For OPARIS we give a high-level description of the algorithms and heuristics used, and some experimental results on real-world problem instances obtained from ISI.