基于蚁群节点寻优的贝叶斯网络结构算法研究

发布时间：2018-04-09 08:51

本文选题：互信息　切入点：蚁群优化　出处：《仪器仪表学报》2017年01期

【摘要】：K2算法是学习贝叶斯网络结构的经典算法。针对K2算法依赖最大父节点数和节点序的不足,以及蚁群算法搜索空间庞大的问题,提出了一种新的贝叶斯结构学习算法-MWST-ACO-K2算法。该算法通过计算互信息建立最大支撑树(MWST),得到最大父节点数;然后利用蚁群算法(ACO)搜索最大支撑树,获得节点顺序;最后结合K2算法得到最优的贝叶斯网络结构。仿真实验结果表明,该方法不仅解决了K2算法依赖先验知识的问题,而且减少了蚁群算法的搜索空间,简化了搜索机制,得到较好的贝叶斯结构。最后将该算法应用到冀东水泥回转窑的实际数据中,构建水泥回转窑的贝叶斯网络结构,提高了故障诊断的准确率。
[Abstract]:K 2 algorithm is a classical algorithm for learning Bayesian network structure.A new Bayesian structure learning algorithm, -MWST-ACO-K2 algorithm, is proposed to solve the problem that K2 algorithm depends on the maximum number of parent nodes and the order of nodes, as well as the huge search space of ant colony algorithm.In this algorithm, the maximum support tree is established by computing mutual information and the maximum number of parent nodes is obtained. Then the maximum support tree is searched by ant colony algorithm (ACO) and the node order is obtained. Finally, the optimal Bayesian network structure is obtained by using K2 algorithm.Simulation results show that this method not only solves the problem that K2 algorithm depends on prior knowledge, but also reduces the search space of ant colony algorithm, simplifies the search mechanism and obtains a better Bayesian structure.Finally, the algorithm is applied to the actual data of cement rotary kiln in Jidong, and the Bayesian network structure of cement rotary kiln is constructed to improve the accuracy of fault diagnosis.
【作者单位】：燕山大学信息科学与工程学院;燕山大学电气工程学院;
【基金】：国家自然科学基金(51641609) 河北省自然科学基金(F2016203354)项目资助
【分类号】：TP18

【参考文献】