[1]张万鹏,王楠,陈璟.一种利用工作流模型的分层任务网络规划领域建模方法[J].智能系统学报,2011,6(3):239-247.
ZHANG Wanpeng,WANG Nan,CHEN Jing.An HTN planning domain modeling method using a workflow model[J].CAAI Transactions on Intelligent Systems,2011,6(3):239-247.
点击复制
ZHANG Wanpeng,WANG Nan,CHEN Jing.An HTN planning domain modeling method using a workflow model[J].CAAI Transactions on Intelligent Systems,2011,6(3):239-247.
一种利用工作流模型的分层任务网络规划领域建模方法
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
6
期数:
2011年第3期
页码:
239-247
栏目:
学术论文—人工智能基础
出版日期:
2011-06-25
- Title:
- An HTN planning domain modeling method using a workflow model
- 文章编号:
- 1673-4785(2011)03-0239-09
- 分类号:
- TP391
- 文献标志码:
- A
- 摘要:
- 为了有效地获取和利用领域知识，提高规划效率，分析了工作流模型和分层任务网络(HTN)规划领域模型的相似性，提出了一种采用工作流模型进行规划领域建模，对领域知识进行获取和表达的方法.工作流模型中的行动和工作流模式，转换为HTN规划中的行动和任务分解；另外，引入了循环(Loop)工作流模式，转换为HTN规划中的递归调用，扩展了工作流模式对规划领域知识的表达能力.在典型的几个规划领域中，引入领域知识后大大提高了规划器的求解效率，从而验证了应用工作流模型进行规划领域建模的有效性.
- Abstract:
- In order to improve planning efficiency by acquiring and utilizing domain knowledge, workflow process models were compared along with hierarchical task network (HTN) planning domain models, and a new method was presented based on the workflow to model planning domain. Activities and workflow patterns were respectively translated into actions and composed tasks in the HTN domain. Loop pattern was introduced to represent recursion in the HTN domain in order to extend the knowledge acquisition and representation abilities of workflow patterns. In experiments on several classical planning domains, such knowledge acquired in workflow models was shown to speed up the classical planner by several orders of magnitude. The experiments validate the HTN planning domain modeling method using a workflow model.
- 参考文献/References:
-
［1］GHALLAB M, NAU D, TRAVERSO P. Automated panning： theory and pactice［M］. San Francisco, USA: Morgan Kaufmann, 2004: 229259.
?［2］KEARNEY P, BORRAJO D, CESTA A, et al. PLANET workflow management R&D roadmap［EB/OL］. (20030930)［20101115］. http://scalab.uc3m.es/~dborrajo/planet/wmtcu/WMTCURoadmapv4CompleteDocument.pdf.
［3］RMORENO M D, BORRAJO D, CESTA A, et al. Integrating planning and scheduling in workflow domains［J］. Expert Systems with Applications, 2007, 33(2): 389406.
［4］ GONZALEZFERRER A, FDEZOLIVARES J, CASTILLO L, et al. Towards the use of XPDL as planning and scheduling modeling tool: the workflow patterns approach［C］//Proceedings of the 11th IberoAmerican Cnference on AI: Advances in Artificial Intelligence. Berlin, Germany: Springer, 2008: 5261.
［5］ GONZALEZFERRER A, FERNANDEZOLIVARES J, CASTILLO L. JABBAH: a Java application framework for the translation between business process models and HTN［C］//The 3rd International Competition on Knowledge Engineering for Planning and Scheduling (ICKEPS 2009). Thessaloniki, Greece, 2009: 2837.
［6］ Workflow Management Coalition. Process definition interface—XML process definition language［EB/OL］. ［20101115］. http://www.wfmc.org/xpdl.html.
［7］ Van DER AALST W M P, TER HOFSTEDE A H M, KIEPUSZEWSKI B, et al. Workflow patterns［J］. Distributed and Parallel Databases, 2003, 14(1): 551.
［8］ RUSSELL S, NORVING P. Artificial intelligence: a modern approach［M］. Upper Saddle River, USA: Pearson Education, 2003: 427428.
［9］ SHUKLA D, SCHMIDT B. Essential Windows workflow foundation［M］. Boston, USA: Pearson Education, 2006: 261292.
［10］ HOFFMANN J, NEBEL B. The FF planning system: fast plan generation through heuristic search［J］. Journal of Artificial Intelligence Research, 2001, 14: 253302.
［11］ALFORD R, KUTER U, NAU D. Translating HTNs to PDDL: a small amount of domain knowledge can go a long way［C］//International Joint Conferences on Artificial Intelligence (IJCAI). Pasadena, USA, 2009: 16291634.
- 相似文献/References:
-
[1]张鸿皓,姜大鹏,庞永杰,等.水下机器人任务流程建模与管理[J].智能系统学报,2013,8(5):433.[doi:10.3969/j.issn.1673-4785.201210067]
ZHANG Honghao,JIANG Dapeng,PANG Yongjie,et al.Workflow modeling and management on AUV[J].CAAI Transactions on Intelligent Systems,2013,8():433.[doi:10.3969/j.issn.1673-4785.201210067]
备注/Memo
收稿日期：2010-11-17.
基金项目：国家自然科学基金资助项目(61005077).
通信作者：张万鹏.E-mail:wpzhang@nudt.edu.cn.
作者简介：
张万鹏，男，1981年生，博士研究生，主要研究方向为智能规划.发表学术论文近10篇.
王楠，男，1981年生，博士研究生，主要研究方向为智能优化.发表学术论文10余篇.
陈璟，男，1972年生，副教授，博士，主要研究方向为人工智能与飞行器规划.发表学术论文20余篇.
更新日期/Last Update:
2011-07-23