摘要
变结构离散动态贝叶斯网络及其推理算法解决了对突变过程的建模和定性推理问题,但是环境突变是随时发生而且无法预计,由此网络结构发生变化后,网络参数必须自适应产生。针对此问题,依据贝叶斯网络的原理,定义了相关节点和最偏好状态的概念,提出了变结构离散动态贝叶斯网络参数的自适应产生算法,并将此算法应用于解决飞行器突发威胁情况下的航路选择问题,取得了满意的结果,验证了所提出的变结构离散动态贝叶斯网络参数的自适应产生算法是可行的,有望解决变结构离散动态贝叶斯网络参数的自适应产生问题。
Structure varied discrete dynamic Bayesian network and its inference algorithm can model the sudden change system, but the change occurred unexpected, when the structure of the discrete dynamic Bayesian networkvaried, the parameters of the discrete dynamic Bayesian network will change consequently. Aimed on this problem, according to the principle of Bayesian network, defined the concepts of relation nodes and prefer status, proposed a self-adaptive algorithm to generate the discrete dynamic Bayesian network parameters when its structure changes, and tested the self-adaptive algorithm with a problem of path selection for the flying vehicle when the threaten appeared suddenly using. The test confirmed the validity of the self-adaptive algorithm to generate the discrete dynamic Bayesian network parameters when its structure changes and implied, the self-adaptive algorithm parameter generating is hopeful to generate the parameters of the discrete dynamic Bayesian network effectively.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2008年第10期1836-1839,共4页
Systems Engineering and Electronics
基金
国家自然科学基金(90205019
60774064)
航空支撑基金(04C53009)资助课题
关键词
变结构
贝叶斯网络
自适应
参数
structure varied
Bayesian network
self-adaptive
parameter
