Exploring the Common Ground between Action Theories and Planning Research Interest.pdfVIP

Explor ing the Common Ground b etween Action Theor ie s and Planning Re s e arch Intere st Yan Zhang Department of Computing University of Western Sydney Nep ean Kingswo o d NSW Australia Email yanstnep eanuwseduau Intro duction Since it was rst p ointed out by McCarthy and Hayes that the frame problem was an obstacle to the formalization of dynamic world mo delling reasoning ab out action has b een widely studied in AI community A numb er of theories and metho dologies for reasoning ab out action have b een prop osed to attack the frame problem and other related problems such as ramica tion and qualication problems So far some imp ortant issues of reasoning ab out action have b een identied and well understo o d by researchers eg the temp oral pro jection problem while other issues are still b eing widely investigated eg indeterminate actions causality in actions concurrent ac tions natural actions etc One common purp ose for the research on action theory is to formalize the pro cedure of reasoning and representing the ef fect of actions in a more realistic environment from b oth foundational and computational asp ects On the other hand planning is designing the b ehavior of some entity that acts either an individual a group or an organization The output usually is a sequence of actions which we call a plan It is obvious that there are closed connections b etween reasoning ab out actions and planning For instance it is essential to formalize the representation of actions b efore a planner can generate a plan for some goal The classical planning syste