摘要
BDD法可以有效地降低故障树分析的代价,其关键就是获取底事件的最优指标顺序,而这又可转化为从六种故障树底事件遍历方式中选择出最佳的问题。文中建立了求解最佳故障树遍历方式的神经网络模型,提出了修正学习率的改进型BP算法。通过利用MATLAB仿真验证了所提模型和算法的有效性。最后分析了各输入节点的相对重要度。
The price of the fault-tree analysis can be effectively reduced by the method of Binary Decision Diagrams (BDD). The ordering of the basic events is critical to the resulting size of the BUD, and ultimately affects the performance and benefits of this technique, so this is an optimal problem for choosing the traversal approach of fault tree. The Neural Network (NN) for choosing optimal traversal approach is modeled in this paper. The improvement of learning rate of error back-propagation algorithm is presented. The model and algorithm is verified by MATLAB simulation. Finally, the relative importance of the input nodes is analyzed.
出处
《弹箭与制导学报》
CSCD
北大核心
2006年第S6期455-457,462,共4页
Journal of Projectiles,Rockets,Missiles and Guidance
基金
西北工业大学青年科技创新基金
西北工业大学研究生创业种子基金(Z20040002)
关键词
故障树
二元决策图
神经网络
BP算法
Fault Trees Binary Decision Diagrams
Neural Networks Back-Propagation algorithm
