Knowledge, action, and the frame problem

AbstractThis paper proposes a method for handling the frame problem for knowledge-producing actions. An example of a knowledge-producing action is a sensing operation performed by a robot to determine whether or not there is an object of a particular shape within its grasp. The work is an extension of Reiter's approach to the frame problem for ordinary actions and Moore's work on knowledge and action. The properties of our specification are that knowledge-producing actions do not affect fluents other than the knowledge fluent, and actions that are not knowledge-producing only affect the knowledge fluent as appropriate. In addition, memory emerges as a side-effect: if something is known in a certain situation, it remains known at successor situations, unless something relevant has changed. Also, it will be shown that a form of regression examined by Reiter for reducing reasoning about future situations to reasoning about the initial situation now also applies to knowledge-producing actions