Abstract

This paper makes the following two contributions to formal theories of actions: Showing that a causal minimization framework can be used effectively to specify the effects of indeterminate actions; and showing that for certain classes of such actions, regression, an effective computational mechanism, can be used to reason about them. Introduction Much recent work on theories of actions has concentrated on primitive, determinate actions. In this paper, we pose ourselves the problem of specifying directly the effects of indeterminate actions, 1 like we do for the primitive, determinate ones. There are several reasons why we think this is an important problem. First of all, there are actions whose effects, when described at a natural level, are indeterminate. Secondly, one can argue that there is no absolute defining line between determinate and indeterminate actions. The differences have a lot to do with the levels of descriptions. The effects of an action may be determinate at one l...

Citations

1359	Some philosophical problems from the standpoint of artificial intelligence - McCarthy, Hayes - 1969
519	The frame problem in the situation calculus: A simple solution (sometimes) and a completeness result for goal regression - Reiter - 1991
300	Circumscription - Lifschitz - 1991
216	State constraints revisited - Lin, Reiter - 1994
166	Embracing causality in specifying the indirect effects of actions - Lin - 1985
104	Achieving several goals simultaneously - Waldinger - 1981
97	Actions with indirect effects: Preliminary report - Kartha, Lifschitz - 1994
92	Synthesizing plans that contain actions with context-dependent effects. Computational Intelligence 4 - Pednault - 1988
64	F.: Forget it - Reiter, Lin - 1994
16	Nondeterministic actions and the frame problem - Boutilier, Friedman - 1995
3	Reasoning About Persistence: A Unified Principle for State Change - Zhang - 1994
2	Reasoning about actions: nondeterministic effects, constraints, and qualification - Barel - 1995