Propositional Planning

We describe a general setting for finite-state planning, where planning operators, or actions, act as functions mapping sets of states to sets of states. In particular, we introduce propositional actions, which generalise STRIPS actions. Sets of states may be represented by Binary Decision Diagrams (BDDs), and propositional actions may be directly encoded as efficient operations on BDDs. We describe PropPlan a planner based on this representation, report its performance on a variety of benchmark problems, and discuss its relation to other BDD-based approaches to planning, and to GraphPlan, by which it was inspired.