基于μ演算的认知难题符号化模型检测
发布时间：2019-05-11 06:24
【摘要】：动态认知逻辑(DEL)是一种研究智能体认知状态变化的一般逻辑方法,不仅可以用于推理多智能体系统(MAS)中的静态认知性质,还可用于推理包含知识更新的MAS系统中的动态认知性质。动态认知逻辑已在认知难题求解、认知规划、安全通信协议、博弈论等多智能体系统研究领域得到越来越深入的应用。本文面向一类多智能体认知难题,提出并实现一种扩展认知计算的μ演算逻辑及其符号化模型检测算法,实验结果表明我们方法的性能优势显著。本文研究成果概括如下:首先设计一种建模描述语言,用于刻画具有线性认知公告行为的认知难题;提出一种融合状态迁移关系和智能体认知关系的认知公告形式模型;设计并实现了一个模型构造算法,将含有认知公告行为的建模描述语言自动转化为相应的认知公告模型;通过在标准μ演算逻辑上扩展认知算子,提出一种新的认知μ演算逻辑,并在认知公告模型上提出了认知μ演算逻辑语义;设计并实现基于有序二元决策图OBDD的认知μ演算符号化模型检测算法;成功地对泥孩子、和与积这两个MAS的经典认知难题进行建模、求解、以及相关时态认知性质的验证。本文研究成果融合了μ演算、静态认知、认知公告(一种动态认知逻辑)的建模与验证方法。所提出的认知μ演算的时态表达能力不仅强于目前主流的时态认知模型检测工具MCK、MCMAS和MCTK,而且也是动态认知模型检测工具DEMO不具备的。上述两个认知难题的实验表明,本文方法的求解效率指数级优于基于DEMO的方法。
[Abstract]:Dynamic cognitive logic (DEL) is a general logical method to study the change of agent cognitive state, which can not only be used to reason the static cognitive properties of multi-agent system (MAS). It can also be used to infer dynamic cognitive properties in MAS systems containing knowledge updates. Dynamic cognitive logic has been more and more deeply applied in the research fields of cognitive problem solving, cognitive planning, secure communication protocol, game theory and so on. In this paper, aiming at a class of multi-agent cognitive problems, a 渭 arithmetic logic and its symbolic model detection algorithm for extended cognitive computing are proposed and implemented. The experimental results show that our method has significant performance advantages. The research results of this paper are summarized as follows: firstly, a modeling description language is designed to describe the cognitive problems with linear cognitive announcement behavior, and a cognitive announcement formal model combining state transition relationship and agent cognitive relationship is proposed. A model construction algorithm is designed and implemented, which automatically converts the modeling description language with cognitive announcement behavior into the corresponding cognitive announcement model. By extending the cognitive operator in the standard 渭 logic, a new cognitive 渭 logic is proposed, and the cognitive 渭 logic semantics is proposed on the cognitive announcement model. The symbolic model detection algorithm of cognitive 渭 calculation based on ordered binary decision graph OBDD is designed and implemented, and the classical cognitive problems of mud child, sum and product are successfully modeled, solved and verified by the relevant temporal cognitive properties. The research results of this paper combine the modeling and verification methods of 渭 calculus, static cognition and cognitive announcement (a dynamic cognitive logic). The temporal expression ability of the proposed cognitive 渭 calculus is not only stronger than the current mainstream temporal cognitive model detection tools MCK,MCMAS and MCTK, but also not available to the dynamic cognitive model detection tool DEMO. The experiments of the above two cognitive problems show that the proposed method is superior to the DEMO-based method in solving the efficiency index.
【学位授予单位】：华侨大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18
本文编号：2474315
[Abstract]:Dynamic cognitive logic (DEL) is a general logical method to study the change of agent cognitive state, which can not only be used to reason the static cognitive properties of multi-agent system (MAS). It can also be used to infer dynamic cognitive properties in MAS systems containing knowledge updates. Dynamic cognitive logic has been more and more deeply applied in the research fields of cognitive problem solving, cognitive planning, secure communication protocol, game theory and so on. In this paper, aiming at a class of multi-agent cognitive problems, a 渭 arithmetic logic and its symbolic model detection algorithm for extended cognitive computing are proposed and implemented. The experimental results show that our method has significant performance advantages. The research results of this paper are summarized as follows: firstly, a modeling description language is designed to describe the cognitive problems with linear cognitive announcement behavior, and a cognitive announcement formal model combining state transition relationship and agent cognitive relationship is proposed. A model construction algorithm is designed and implemented, which automatically converts the modeling description language with cognitive announcement behavior into the corresponding cognitive announcement model. By extending the cognitive operator in the standard 渭 logic, a new cognitive 渭 logic is proposed, and the cognitive 渭 logic semantics is proposed on the cognitive announcement model. The symbolic model detection algorithm of cognitive 渭 calculation based on ordered binary decision graph OBDD is designed and implemented, and the classical cognitive problems of mud child, sum and product are successfully modeled, solved and verified by the relevant temporal cognitive properties. The research results of this paper combine the modeling and verification methods of 渭 calculus, static cognition and cognitive announcement (a dynamic cognitive logic). The temporal expression ability of the proposed cognitive 渭 calculus is not only stronger than the current mainstream temporal cognitive model detection tools MCK,MCMAS and MCTK, but also not available to the dynamic cognitive model detection tool DEMO. The experiments of the above two cognitive problems show that the proposed method is superior to the DEMO-based method in solving the efficiency index.
【学位授予单位】：华侨大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18
【参考文献】
相关期刊论文 前5条
1 骆翔宇;谭征;苏开乐;吴立军;;一种基于认知模型检测的Web服务组合验证方法[J];计算机学报;2011年06期
2 骆翔宇;苏开乐;顾明;;一种求解认知难题的模型检测方法[J];计算机学报;2010年03期
3 郭美云;;从PAL看认知逻辑的动态转换[J];自然辩证法研究;2006年01期
4 苏开乐,骆翔宇,吕关锋;符号化模型检测CTL[J];计算机学报;2005年11期
5 刘瑞胜,孙吉贵,刘叙华;认识逻辑(1):关于知识和信念的逻辑框架[J];计算机学报;1998年07期
相关博士学位论文 前4条
1 吴梅红;基于动态认知逻辑的多主体系统知识推理研究[D];厦门大学;2009年
2 骆翔宇;多智能体系统的符号模型检测[D];中山大学;2006年
3 郭美云;带有群体知识的动态认知逻辑[D];北京大学;2006年
4 于江涛;多智能体模型、学习和协作研究与应用[D];浙江大学;2003年
，本文编号：2474315
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2474315.html
