基于BP神经网络的某反装甲武器系统打击效果预测

Prediction of Strike Effect of an Anti-Armor Weapon System Based on BP Neural Network

针对实装试验难以实现某型反装甲武器系统全样本条件下的打击效果预测,提出一种基于梯度搜索技术的改进BP神经网络的反装甲武器系统打击效果模型,建立影响预测结果的影响因子量化值和目标靶车打击效果的量化值之间的对应关系,利用作战试验历史数据作为样本训练模型,最后以某次真实试验结果检验模型。结果表明,打击效果预测值与真实值吻合较好,BP神经网络可实现对该武器系统对靶车打击效果进行有效预测,可为火力打击方案制订、作战试验方案评估、训练效果评估等提供决策依据。

Aiming at that the anti-impact effect prediction of a certain type of anti-armor weapon system under full sample condition is difficult to implement,an improved BP neural network based on gradient search technology was proposed to predict the impact effect of anti-armor weapon system.The BP neural network model established the correspondence between the impact factors and the quantified values of strike effect of the target tank.We used the historical data of the combat test as the sample training model.Finally,we test the predicted results with a real test data result.The results show that the predicted value of the strike effect is in good agreement with the true value.The BP neural network can effectively predict the strike effect of the weapon system attack target.This method can provide decision-making basis for firepower strike program formulation,evaluation of combat test plan,and evaluation of training effects.