一种求解认知难题的模型检测方法  

A Model Checking Approach for Solving Epistemic Riddles

作　　者：骆翔宇[1,2] 苏开乐[3] 顾明[1] 

机构地区：[1]清华大学软件学院信息系统安全教育部重点实验室,北京100084 [2]桂林电子科技大学计算机与控制学院,广西桂林541004 [3]北京大学信息科学技术学院,北京100871

出　　处：《计算机学报》2010年第3期406-414,共9页Chinese Journal of Computers

基　　金：国家自然科学基金重点项目(90718039);国家"九七三"重点基础研究发展规划项目基金(2010CB328103);国家杰出青年科学基金(60725207);国家自然科学基金(60763004);广西青年科学基金(桂科青0728090);中国博士后科学基金(20090450389)资助

摘　　要：用公告逻辑建模并求解和与积认知难题.提出一种动态认知模型,将环境认知模型与公告导致的认知模型线性组合,从而在时态认知逻辑模型检测技术中扩展支持公告逻辑的建模与验证.该模型检测方法不仅可以用于搜索认知难题的所有解,而且可以验证相关的时态认知性质,这一特性是当前认知逻辑模型检测工具MCK、MCMAS和DEMO不能完全支持的.作者采用OBDD开发了相关的符号化模型检测工具MCTK并对和与积难题进行建模和验证,实验结果说明了文中方法的正确性和高效性.This paper aims to model and solve the Sum and Product Riddle in public announcement logic.A dynamic epistemic model is proposed,that is the linear temporal combination of the epistemic model of environment and the epistemic models after each announcement,such that the authors＇ model checking technique for temporal epistemic logic can be extended to support the modeling and verification of public announcement logic.This model checking method not only can help to find all solutions,but also verify related temporal epistemic properties.Such characteristic is not fully supported by the current version of MCK,MCMAS and DEMO.The authors implemented the proposed method in the symbolic model checker MCTK via OBDD and verified the sum and product riddle.The experimental results show that the proposed method is correct and efficient.

关 键 词：模型检测 OBDD 公告逻辑 时态认知逻辑 和与积难题 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]