M. Ginsberg, H. Holbrook. What defaults can do that hierarchies can't. Fundamental Informaticae (21) (1994), pages 149-- 159.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....its role by the player number given from server. Behaviours according to the roles are due to special plans (Goal Keeper, Defender, Attacker, etc. Players can change their roles depending on the situation on the field. 4 Conclusion Our approach is somewhat analogous to that presented in [2, 3, 4], with one essential difference: our planning system works on line, and plans refinements are being made dynamically in case of need. More detailed comparisons deserve the further investigations and are to be presented elsewhere. The further development of the method itself and the more precise ....

M. Ginsberg, H. Holbrook. What defaults can do that hierarchies can't. Fundamental Informaticae (21) (1994), pages 149-- 159.

....that leads to approximate plans. The notion of an approximate plan is reminiscient of Elkan s use of approximate plans in his Perflog calculus. However, he follows a one shot approach where information gained during an interrupted reasoning process cannot be reused. Ginsberg and Holbrook [9] also make the distinction between strict preconditions and default preconditions (in action precondition axioms) They show that the distinction lends itself to a natural assignment of priorities to goals which is important for hierarchical planning; strict conditions have high priority while ....

Matthew L. Ginsberg and Hugh W. Holbrook. What defaults can do that hierarchies can't. In Proceedings of the 1992 Nonmonotonic Resoning Workshop, 1992.

....assumption about the world. Since assumptions are defeasible, actual systems must monitor the execution of the plan, and replan if necessary [FHN72, Wil88] Certain planners can be viewed as making assumptions (which are later dropped) in order to quickly derive an abstract or approximate plan [Sac74, Elk90, GH92]. In this paper, we ignore the computational benefits of making assumptions, and focus on the issue of constructing assumptions that induce the agent to construct a better plan. In order to determine whether an assumption results in a better plan, we need to judge when one plan is better than ....

....made was wrong. In this case, the agent should stop and replan, using a new set of assumptions [FHN72, Wil88] The process of replacing an assumption with a new one is closely related to belief revision. The connection between assumptions and defaults is not limited to this aspect. For example, in [Elk90, GH92], defaults are used in the planning process in order to focus the search on the difficult aspects of the plan. That is, there is a default assumption that certain subgoals, that are relatively easy to achieve, are in fact achievable. Our framework does not address this use of assumptions. However, ....

M. L. Ginsberg and H. W. Holbrook. What defaults can do that hierarchies can't. In Proceedings 1992 Nonmonotonic Reasoning Workshop, Plymouth, VT, 1992.

....a balloon causing a noise) 8] or have ramifications that need to be computed based on the values of other fluents [3, 13] All of this work remains in the same overall declarative framework and continues to use this framework to provide the machinery needed for planning. As discussed elsewhere [12], we can also introduce defaults into the declarative language, thereby capturing in our setting the ideas typically associated with hierarchical planners. 5.3 Status The current implementation is a good but not perfect match for the theoretical constructs that we have discussed. Plans and ....

M. L. Ginsberg and H. W. Holbrook. What defaults can do that hierarchies can't. In Proceedings 1992 Nonmonotonic Reasoning Workshop, Plymouth, VT, 1992.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC