Handling Implication and Universal Quantification Constraints in FLUX

FLUX is a CLP-approach for programming agents that reason about actions under incomplete state knowledge. FLUX is based on the solution to the fundamental frame problem in the fluent calculus. The core is a set of Constraint Handling Rules for the constraints that are used to encode state knowledge. In order to allow for efficient constraint solving, the original expressiveness of state representations in FLUX has been carefully restricted. In this paper, we enhance the expressiveness by adding both implication and universal quantification constraints. We do so without losing the computational merits of FLUX. We present a set of Constraint Handling Rules for these new constraints and prove their correctness against the fluent calculus.