神经网络和DEA方法的炮兵火力毁伤先验评估模型

摘要

The model of prior evaluation of artillery firepower bombardment is studied by the methods of B-P Neural Network and DEA (Data Envelopment Analyse). The set of samples based on the experience data is set up and the artillery firepower bombardment plan is analyzed by the B-P Neural Network prediction model and the damage benefit is forecasted. Then the DEA evaluated model with the preference cone structure is established and the rationality and feasibility of the firepower plan. At last,it the example analysis of the 17 group's artillery firepower damage data is given. The result indicates that the preevaluation result given by this model is scientific,clear and credible.%提出运用BP神经网络理论和数据包络分析(DEA)方法,对炮兵火力毁伤先验评估问题进行研究.利用以往经验数据建立样本集,通过BP神经网络预测模型,对待评估的炮兵火力毁伤计划进行分析,预测其毁伤效益,再建立带有偏好锥结构的DEA评估模型,分析火力计划的可行性和合理性；最后以某部17组炮兵火力毁伤数据进行实例分析.结果表明,该模型能给出科学、明确、可靠的先验评估结果.