基于BAM-Hamming神经网络的风险估计研究

现代弹药系统是一种集多门学科于一身的复杂系统 ,在研制过程中 ,由于受到不确定性因素的影响 ,会给研制工作带来一定的风险。提出了风险分析在弹药系统研制中的重要性 ,分析了各种风险分析与估计方法 ,给出了项目研制中风险特征因子及其估计方法 ,并将其转换成相应的二元码序列。利用BAM Hamming神经网络建立风险估计模型 ,采用最优编码方式 ,提高神经网络的容错性。以反直升机智能雷为例 ,验证了该方法的有效性和可靠性 ,以及一定的实用价值。

Study on Risk Estimation Based on the BAM-Hamming Neural Network

Modern ammunition system is a complicated system involving multiple disciplines. During its development, the presence of voice uncertain factors, brings many risks to the development work. In this paper the importance of risk analysis to the ammunition system development is elaborated and various methods of risk analysis and risk estimation are studied. The risk characteristic factors in the project development and their estimation method are given, and then the characteristic factors are transformed into the corresponding binary code serials. During the establishment of the risk estimation model, the BAM Hamming neural network is used, and the optimum encoding method is adopted, thus the error tolerance of the neural network is improved. With anti helicopter intelligent mine as an example, the efficiency and reliability as well as the practical value of the method are verified.