Automated Planning has been one of the traditional strengths of University of Washington Artificial Intelligence group, which has been responsible for making steady contributions in the field for the last 18 years. Here is a list of the group's projects (in roughly reverse chronological order):

Current Projects

• MDP Solvers That Scale: Developing algorithms for solving large MDPs.
• The theory of temportal planning: We are characterizing what aspects of a planning domain (or modeling language) makes temporal planning difficult.

• Concurrency in MDPs: Augmenting Markov Decision Processes to handle concurrent temporally-extended actions.
• BlackBox: A planning system that combines SAT technology with Graphplan.
• TGP: A fast temporal planner, aimed at NASA spacecraft domains.
• SGP: A planner that handles uncertainty, sensory actions and combines conformant and contingent planning.
• LPSAT: A metric resource planner, which compiles planning problems to a combination of linear inequalities and SAT formulas --- to be solved by a novel hybrid solver, combining a systematic satisfiability solver with an incremental Simplex algorithm.
• Theory of Planning Under Uncertainty: Investigating the computability of Markov Decision Processes.
• MEDIC: Compiles STRIPS problems into satisfiability problems.
• PYRRHUS: Finds optimal plans for goal-directed value functions.
• ZENO: Temporal planner that handles deadline goals and continuous change.
• BURIDAN: Probabilistic planner reasons about uncertainty.
• CBURIDAN: Extends BURIDAN with sensing actions and contingent execution.
• XII: Executes sensing actions to handle incomplete information with the Internet softbot.
• OCCAM: Planner optimized for information gathering and controls the Razor softbot.
• UCPOP: Once the standard benchmark planner (distributed to 100+ sites), UCPOP handled universal quantification, conditional effects, and dynamic universes with a sound and complete algorithm.
• FABIAN: Plans with abstract actions which it automatically generates.