APPSSAT: Approximate Probabilistic Planning Using Stochastic Satisfiability [chapter]

Stephen M. Majercik
2005 Lecture Notes in Computer Science  
We describe APPSSAT, an anytime probabilistic contingent planner based on ZANDER, a probabilistic contingent planner that operates by converting the planning problem to a stochastic satisfiability (Ssat) problem and solving that problem instead [1] . The values of some of the variables in an Ssat instance are probabilistically determined; APPSSAT considers the most likely instantiations of these variables (the most probable situations facing the agent) and attempts to construct an approximation
more » ... of the optimal plan that succeeds under those circumstances, improving that plan as time permits. Given more time, less likely instantiations/situations are considered and the plan is revised as necessary. In some cases, a plan constructed to address a relatively low percentage of possible situations will succeed for situations not explicitly considered as well, and may return an optimal or near-optimal plan. We describe experimental results showing that APPSSAT can find suboptimal plans in cases in which ZANDER is unable to find the optimal (or any) plan. Although the test problems are small, the anytime quality of APPSSAT means that it has the potential to efficiently derive suboptimal plans in larger, time-critical domains in which ZANDER might not have sufficient time to calculate any plan. We also suggest further work needed to bring APPSSAT closer to attacking real-world problems.
doi:10.1007/11518655_19 fatcat:zej4ze2yzvajhj5burprwjjqpa