Decision-Theoretic Subgoaling for Planning with External Events
Jim Blythe · 1994
I describe a planning methodology for domains with uncertainty in the form of external events that are not completely predictable. Under certain conditions, these events can be modelled as continuous-time Markov chains whose states are characterised by the planner's domain predicates. Planning is goal-directed, but the subgoals are suggested by analysing the utility of the partial plan rather than being simply the open conditions of the operators in the plan, a technique I call "decision-theoretic subgoaling ". Other planners for uncertain domains can be viewed as performing decision-theoretic subgoaling, which I argue is a useful way to combine AI-based planning and decision theory 1 . 1 Introduction One of the central assumptions of classical planning is that the state resulting at some time after performing an action can be predicted completely and with certainty. This assumption permits a style of planning in which a goal, represented by a sentence in first-order ...