改进BP神经网络的空地制导弹药打击目标价值评估

Air-to-ground Guided Munitions Attacking Target Value Evaluation Based on Improved BP Neural Network

目标价值评估是空地制导弹药打击行动中重要的一环。结合空地制导弹药作战的特点选取了6个目标价值评估指标;利用贝叶斯正则化方法和遗传算法优化BP神经网络,建立了目标价值评估模型;对模型进行了训练和测试,测试结果表明,该模型评估精度高、泛化能力强,能够有效帮助指挥机关评估打击目标价值,为决策提供依据或参考。

Target valve evaluation is an important part of air-to-ground guided munitions action. Six indexes are selected to evaluate target value,combined with the characteristic of air-to-ground guided munitions. A model of target value evaluation with BP neural network is established,which is optimized by Bayesian regularization method and genetic algorithm. Finally,the model is trained and tested. The result shows that the model owns high evaluation precision and good generalization performance,can effectively help command organs evaluate the target value and provides the basis and reference for decision-making.