变结构离散动态贝叶斯网络的快速推理算法  

To Speed up the Referrence for the Structure Varied Discret Dynamic Bayesian Networks

作　　者：史建国[1] 高晓光[2] 

机构地区：[1]军航空工程学院,山东烟台264001 [2]西北工业大学电子工程学院,西安710072

出　　处：《火力与指挥控制》2012年第9期71-74,共4页Fire Control & Command Control

基　　金：国家自然科学基金(60774064);博士后研究基金资助项目

摘　　要：变结构离散动态贝叶斯网络是对传统离散动态贝叶斯网络的推广,具有更广泛的建模应用价值,但是其推理算法还有待进一步完善。针对变结构离散动态贝叶斯网络的推理算法难以理解、编程计算难、推理速度慢的问题,给出了实现变结构离散动态贝叶斯推理算法的数据结构,并推导了进行并行计算的推理算法和编程步骤,并通过实例进行了算理验证。给出的方法对变结构离散动态贝叶斯网络的编程应用具有参考价值,同时可以加快变结构离散动态贝叶斯网络的推理计算速度。The Structure Varied Discret Dynamic Dynamic Bayesian Network. It can be applyed more Bayesian Network is the widened traditional Discret widened than traditional Discret Dynamic Bayesian Network. But its inference algorithm is not super enough. Aim at the shortcomings of the Structure Varied Discret Dynamic Bayesian Networks,such as its inference algorithm is hard to understand, hard to progamming and running slowly, this paper, proposes the most suitable storage data structure of the Structure Varied Discret Dynamic Bayesian Networks, deduces the parallel inference algorithm for the Structure Varied Discret Dynamic Bayesian Networks, and verifies the parallel inference algorithm for the Structure Varied Discret Dynamic Bayesian Networks by a sample. The parallel inference algorithm is useful for applying the Structure Varied Diseret Dynamic Bayesian Networks, and speed up the inference running of the Structure Varied Diseret Dynamic Bayesian Networks.

关 键 词：贝叶斯网络 算法 数据结构 推理 

分 类 号：TN29[电子电信—物理电子学] TN18