Engineering Algorithms for Workflow Satisfiability Problem with User-Independent Constraints

The workflow satisfiability problem (WSP) is a planning problem. Certain sub-classes of this problem have been shown to be fixed-parameter tractable. In this paper we develop an implementation of an algorithm for WSP that has been shown, in our previous paper, to be fixed-parameter for user-independent constraints. In a set of computational experiments, we compare our algorithm to an encoding of the WSP into a pseudo-Boolean SAT problem solved by the well-known solver SAT4J. Our algorithm solves all instances of WSP generated in our experiments, unlike SAT4J, and it solves many instances faster than SAT4J. For lightly constrained instances, SAT4J usually outperforms our algorithm.