Decision Theory, the Situation Calculus and Conditional Plans

摘要

This paper shows how to combine decision theory and logical representations of actions in a manner that seems natural for both. In particular, we assume an axiomatization of the domain in terms of situation calculus, using what is essentially Reiter's solution to the frame problem, in terms of the completion of the axioms defining the state change. Uncertainty is handled in terms of the independent choice logic, which allows for independent choices and a logic program that gives the consequences of the choices. As part of the consequences are a specification of the utility of (final) states, and how (possibly noisy) sensors depend on the state. The robot adopts conditional plans, similar to the GOLOG programming language. Within this logic, we can define the expected utility of a conditional plan, based on the axiomatization of the actions, the sensors and the utility. Sensors can be noisy and actions can be stochastic. The planning problem is to find the plan with the highest expected utility. This representation is related to recent structured representations for partially observable Markov decision processes (POMDPs); here we use stochastic situation calculus rules to specify the state transition function and the reward/value function. Finally we show that with stochastic frame axioms, action representations in probabilistic STRIPS are exponentially larger than using the representation proposed here. 1 1