动态贝叶斯网络与黑板机制结合的协同决策方法_马国普.pdfVIP

32 12 Vol. 32 No. 12 2010 12 Systems Engineering and Electronics December 2010 : 1001506X(2010) 1225 005 马国普, 沙基昌, 陈俊良, 陈 超, 姜 鑫 ( 国防科学技术大学信息系统与管理学院, 湖南长沙410073) : 针对信息化战争中重大复杂决策问题, 提出动 贝叶斯网络与黑板机制相结合的协同决策方法以 协同防空预警雷达目标识别问题为背景, 针对复杂问题协同决策的特点, 将动 贝叶斯网络的定量推理能力与黑 板机制灵活集成多领域决策者智慧定性推理的能力相结合, 提出动 贝叶斯黑板协同决策机制在概率意义下 证明了该方法的有效性, 并用仿真试验验证该方法可以较显著地提高协同决策的正确性 : 决策; 动 贝叶斯网络; 黑板机制; 协同; 动 贝叶斯黑板 : C 34 : A DOI: 10.3 6 / j. issn. 1001506X.2010. 12. 20 Coordination decisionmaking based on dynamic Bayesian network and blackboard mechanism MA Guopu, SHA Jichang, CHEN Junliang, CHEN Chao, JIANG Xin (Coll. of Inf ormation System and Management, National Univ . of Def ense Technology , Changsha4100 3, China) Abstract: A coordination decisionmaking method is proposed based on dynamic Bayesian network ( DBN) and blackboard mechanism for complex decisionmaking problem in inf ormation war. In the background of the radar target discrimination problem, the quantitative ratiocination capability of DBN and the qualitative ratioci nation capability by integrating wisdoms of heterogeneous experts of blackboard mechanism are integrated to propose Bayesian blackboard mechanism. Then the eff ectiveness of this method is proved by theory analysis, and a simulation experiment is implemented to show that the method can markedly improve the correct rate. Keywords: decisionmaking; dynamic Bayesian network; blackboard mechanism; coordination; dynamic Bayesian blackboard 0 ; , , , , , , ( dynamic Bayesian , netw ork, DBN) ,