基于BP神经网络误差函数优化算法的武器装备承制商信用评价模型研究  

Research on Credit Evaluation Model Design and Simulation of Weapon Equipment Suppliers

作　　者：张蓉[1] 陈云翔[1] 李大伟[1] 

机构地区：[1]空军工程大学装备管理与安全工程学院,陕西西安710038

出　　处：《数学的实践与认识》2012年第21期141-147,共7页Mathematics in Practice and Theory

基　　金：国防基金项目(51327020104)

摘　　要：信用评价是选择武器装备承制商的重要手段.以国标为基础,结合承制商具体情况确定了信用评价指标体系.分析了传统信用评价方法的不足,对经典BP神经网络的误差函数进行优化,优化后的网络模型收敛速度更快,预测精度更高.构建BP神经网络武器装备承制商信用评价模型,仿真实验表明武器装备承制商信用评价可以选用BP神经网络模型.Credit evaluation is an important means to select weapon equipment suppliers. Integrated with characteristics of weapon equipment suppliers, credit evaluation index system is established based on China＇s National Standard. Deficiency of traditional credit evaluation method is analyzed. Error function of traditional BP neural network is improved, which has a faster convergence and higher prediction accuracy. Credit evaluation model of weapon equipment suppliers based on BP neural network is set up. Simulation results show that credit evaluation model of weapon equipment suppliers based on BP neural is valid.

关 键 词：武器装备承制商 信用评价 BP神经网络 误差函数 

分 类 号：E920[军事—军事装备学] O242.1[兵器科学与技术—武器系统与运用工程]