| Nau, D. S., Smith, S. J. J.,Erol, K.: Control Strategies in HTN Planning: Theory versus Practice. In Proc. of AAAI-98/IAAI-98, Madison, WI, AAAI Press, 1127---1133, 1998 |
....in the emergence of story variants of a global theme, which arise from the interactions of virtual characters. As mentioned before, HTNs have been successfully introduced in these domains as an explicit representation for the set of all possible solutions or plans managed by the characters [3][7]. In order to manage the associated knowledge of the storytelling problem, it can be split in independent subtasks that will be managed by HTNs for the necessary decision taking of the virtual actors, which finally create the story. We have presented the MinMin search domain in a similar problem ....
Nau, D.S., Smith, S.J.J., and Erol, K., 1998. Control Strategies in HTN Planning: Theory versus Practice. Proceedings of AAAI/IAAI-98, pp. 1127-1133.
....presented in this paper: subgoals can be interleaved and all linearisations of subgoals are valid. This #exibility preserves the expressiveness of standard HTN planning (which is semi decidable, see [14] while TIGNUM2 s planner is strictly less expressive (EXPSPACE hard and in DOUBLE EXPTIME, [27]) The extra #exibility in GoBI s planner is crucial for games such as Go where interleaving goals is a natural part of play. In Go, capturing a group for example may involve several tasks (surrounding it, preventing eyes etc. which need to be achieved in pseudo parallel. That is, concentrating ....
....5.4 above have signi#cant advantages over those using partial order decompositions. This observation runs contrary to conventional wisdom in the AI planning community. The least commitment aspect of partial order decompositions was always thought to bring signi#cant advantages. Recent work in [27] discusses the assumptions underlying the use of partial order, backwards chaining strategies for HTN planners and suggests that they are not always valid (and in particular do not apply for Bridge play and process planning) In the context of an adversarial planning domain we can elaborate on ....
D.S. Nau, S.J.J. Smith, K. Erol, Control strategies in HTN planning: theory versus practice, Proc. Innovative Applications of Arti#cial Intelligence Conf. (in conjunction with AAAI'98), AAAI Press, 1998, pp. 1127--1133.
....in this paper: subgoals can be interleaved and all linearisations of subgoals are valid 30 . This flexibility preserves the expressiveness of standard HTN planning (which is semi decidable, see [14] while Tignum2 s planner is strictly less expressive (EXPSPACE hard and in DOUBLE EXPTIME, [27]) The extra flexibility in gobi s planner is crucial for games such as Go where interleaving goals is a natural part of play. In Go, capturing a group for example may involve several tasks (surrounding it, preventing eyes etc. which need to be achieved in pseudo parallel. That is, concentrating ....
....x5.4 above have significant advantages over those using partial order decompositions. This observation runs contrary to conventional wisdom in the AI planning community. The least commitment aspect of partial order decompositions was always thought to bring significant advantages. Recent work in [27] discusses the assumptions underlying the use of partial order, backwards chaining strategies for HTN planners and suggests that they are not always valid (and in particular do not apply for Bridge play and process planning) In the context of an adversarial planning domain we can elaborate on ....
D. S. Nau, S. J. J. Smith, and K. Erol. Control Strategies in HTN Planning: Theory Versus Practice. In Proceedings of the Innovative Applications of Artificial Intelligence Conference (in conjunction with AAAI'98), pages 1127-- 1133. AAAI Press, 1998.
....as a set of numeric equations, which should make it quite easy to incorporate numeric constraints and computations into formulations of planning domains. The ability to do this is a critical need for real world planning, and it is not addressed adequately in most existing AI planning systems (Nau et al. 1998). Our work is still in its early stages, and the results reported above are still preliminary. Although the first results are encouraging, our first goal is to investigate whether they generalize to other planning problems (Logistics problem, Rocket problem) Also, we are looking at alternative ....
D. S. Nau, S. J. Smith and Kutluhan Erol. 1998. Control strategies in HTN planning: theory versus practice. In Proc. IAAI-98, to appear.
.... quite naturally allow the incorporation of numeric constraints and objectives into planning domains(for example, see Kautz and Walser [1999] The use of numerical constraints and objectives is not addressed adequately in most existing AI planning systems, but it is critical in real world planning [Nau et al. 1998]. One di#culty indeveloping Integer Programming formulations for AI planning is that the performance of the resulting IP will depend critically on how AI planning problems are formulated as Integer Programs. The purpose of this paper is therefore to develop good domainindependent IP formulations ....
D. S. Nau, S. J. Smith and Kutluhan Erol. 1998. Control strategies in HTN planning: theory versus practice. In AAAI-98/IAAI-98 Proceedings, 1127--1133, 1998.
....IMPACTing SHOP ized Oracle and nested multi record TAADS data [SRM98] a variety of US Army simulation data from a massive program called JANUS, Training and Instrumentation Command, a face recognition program, and so forth. 2. The need to perform mixed symbolic numeric reasoning. For example, NSE98] describes the need to reason about a variety of numeric and symbolic conditions in order to do manufacturing planning and to plan declarer play in the game of bridge. 3. The need to coordinate multiple agents. For example, in planning the movement of a cargo container from its point of origin ....
Dana S. Nau, Stephen J. J. Smith, and Kutluhan Erol. Control Strategies in HTN Planning: Theory versus Practice. In AAAI-98/IAAI-98 Proceedings, pages 1127-1133, 1998.
....Oracle and nested multi record TAADS 1 data [Schafer et al. 1998] a variety of US Army simulation data from a massive program called JANUS, Training and Instrumentation Command, a face recognition program, and so forth. ffl The need to perform mixed symbolic numeric reasoning. For example, [Nau et al. 1998] describes the need to reason about a variety of numeric and symbolic conditions in order to do manufacturing planning and to plan declarer play in the game of bridge. ffl The need to coordinate multiple agents. For example, in planning the movement of a cargo container from its point of origin ....
Nau, D. S., Smith, S. J. J., and Erol, K. (1998). Control Strategies in HTN Planning: Theory versus Practice. In AAAI-98/IAAI-98 Proceedings, pages 1127--1133.
No context found.
D. S. Nau, S. J. Smith and Kutluhan Erol. 1998. Control strategies in HTN planning: theory versus practice. In Proceedings of the 16th National Conference of the American Association for Artificial Intelligence, pp 1127-1133.
No context found.
Nau, D.; Smith, S. J.; and Erol, K. 1998. Control Strategies in HTN Planning: Theory versus Practice. AAAI-98/IAAI-98, 1127--1133.
No context found.
Nau, D. S., Smith, S. J. J.,Erol, K.: Control Strategies in HTN Planning: Theory versus Practice. In Proc. of AAAI-98/IAAI-98, Madison, WI, AAAI Press, 1127---1133, 1998
No context found.
Nau, D.S., Smith, S.J.J., Erol, K.: Control strategies in HTN planning: Theory versus practice. In: AAAI/IAAI. (1998) 1127--1133
No context found.
D. S. Nau, S. J. J. Smith, and K. Erol. Control strategies in HTN planning: Theory versus practice. In AAAI/IAAI, pages 1127--1133, 1998.
No context found.
D.S. Nau, S.J. Smith, and K. Erol, "Control Strategies in HTN Planning: Theory versus Practice," Proc. AAAI/IAAI-98, AAAI Press, Menlo Park, Calif., 1998, pp. 1127--1133.
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