[1]谷文祥,杨永娟,闫书亚.ji基于条件效果的对象动态可变图规划[J].智能系统学报,2007,2(3):12-18.
GU Wen-xiang,YANG Yong-juan,YAN Shu-ya.Creating or deleting objects graphplan based on condition al effects[J].CAAI Transactions on Intelligent Systems,2007,2(3):12-18.
点击复制
GU Wen-xiang,YANG Yong-juan,YAN Shu-ya.Creating or deleting objects graphplan based on condition al effects[J].CAAI Transactions on Intelligent Systems,2007,2(3):12-18.
ji基于条件效果的对象动态可变图规划
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
2
期数:
2007年第3期
页码:
12-18
栏目:
学术论文—人工智能基础
出版日期:
2007-06-25
- Title:
- Creating or deleting objects graphplan based on condition al effects
- 文章编号:
- 1673-4785(2007)03-0012-07
- Keywords:
- graphplan; correlative component; irrelative compon ent; creating or deleting objects component; conditional effects
- 分类号:
- TP18
- 文献标志码:
- A
- 摘要:
- 主要研究了基于条件效果的对象动态可变的规划问题.提出了相关元件、无关元件、创建/删除对象元件和普通元件等概念, 把带有条件效果的动作和不带有条件效果的动作都元件化,并采用了对象命题化的思想.给出了新的基于目标驱动的规划图扩展算法和前向搜索有效规划算法,并给出了相应的后向传播互斥的定义,使得规划图的规模比较小,减少了搜索空间,大大提高了求解有效规划的效率.由于算法中的动作创建的效果是依赖于上下文的描述,这更加符合现实需要,使处理的问题更接近于真实的世界状态,因而此算法比以往的算法应用性更强,更具有现实意义.
- Abstract:
- Mainly research was done on the creating or deleting objects Graphplan that based on conditional effects. Firstly, several new concepts wer e propos ed, such as correlative component, irrelative component, creating or deleting ob jects component, common component and so on. Secondly, both actions with conditi onal effects and that without conditional effects were considered as components and the conception of transforming object into proposition was introduced. Third ly, a novel intelligent planning algorithm which expanded the planning graph bac kwards from the goal set and searched a valid plan forward was proposed and also a new definition of mutex inference backwards was given correspondingly. The me thod reduced the scale of the planning graph and the search space, improved the efficiency of searching the valid plan greatly. Because the effects created by a ctions were contexdependent, it was more suitable for the practical needs comp ared with previous methods, and made the planning problems to be handled much cl oser to the real word. Therefore, the method has its advantage over previous one s in application and has more practical significance.
- 参考文献/References:
-
［1］BLUM A, FURST M. Fast planning through planning graph analysis［A］ . In Proc 14th Int Joint Conf AI［C］. Montreal, Canada, 1995.
［2］BLUM A, FURST M. Fast planning through planning graph analysis［J］. J Ar tificial Intelligence, 1997,90(1-2):281-300.
［3］ANDERSON C R, SMITH D E,WELD D S. Conditional effects in graphplan［A］. In Proc AI Planning Systems Conference［C］. AAAI Press. Melo Park, 1998.
［4］WELD D S, ANDERSON C R, SMITH D E. Extending graphplan to handle uncertain and sensing actions［A］. In AAAI98［C］. Madison, Wisconsin, U S A, 1998.
［5］BLUM A L, LANGFORD J C. Probabilistic planning in the graphplan framework ［A］.AIPS98 Workshop on Planning as Combinatorial Search［C］. Pittsburgh , 199 8.
［6］PARKER E. Making graphplan goalDirected［A］. In ECP99［C］. Durham, Unit ed Kingdom , 1999.
［7］HONG J. Graph construction and analysis as a paradigm for plan recognition ［A］. In Proc of AAAI2000: Seventeenth National Conference on Artificial Inte lligence［C］. Austin, Texas, USA, 2000.
［8］GU Wenxiang, CAI Zengyu, ZHANG Xinmei, et al. Extending graphplan to handl e the creating or destroying objects panning［A］. ICMLC2005［C］. Guangzhou,Chi na, 2005.
［9］NEBEL B, DIMOPOULOS Y,KOEHLER J. Ignoring irrelevant facts and opera tors in plan generation［A］. In Proc 4th European Conference on Planning［C］. Toulouse, France, 1997.
［10］XU Li, GU Wenxiang, ZHANG Xinmei, et al. Goal-directed flexible graphplan ［A］. In Proceeding of 2005 International Conference on Machine Learning and Cy bernetics［C］. Guangzhou, China, 2005.
［11］KAMBHAMPATI S, PARKER E，LAMBRECHT E. Understanding and extending gr aphpla n［A］. In proceedings of the 4th European Conference on Planning Toulouse［C］. France, 1997.
［12］MIGUEL I, JARVIS P，SHEN Q. Flexible graphplan［A］. In Proceedings of the Fourteenth European Conference on Artificial Intelligence［C］. Berlin, Humbold t University, 2000.
［13］MITTAL S，FALKENHAINER B. Dynamic constraint satisfaction problems［A］. I n Proc. of the 8th National Conference on Artificial Intelligence［C］. Boston, Massachusetts, 1990.
备注/Memo
收稿日期：2006-10-25.
作者简介：
谷文祥，男，1947年生，教授，博士生导师，主要研究方向为智能规划与规划识别、形式语言与自动机、模糊数学及其应用.主持国家自然科学基金项目2项，参与国家自然科学基金项目1项，教育部重点项目2项，省科委项目1项，发表论文百篇以上. E-mail:gwx@nenu.edu.cn.
杨永娟，女，1982年生，硕士研究生，主要研究方向为智能规划与规划识别.
闫书亚，女，1982年生，硕士研究生，主要研究方向为智能规划与规划识别.
更新日期/Last Update:
2009-05-06