基于串联QFD神经网络的武器装备使用需求映射方法  

Equipment Application Demand Mapping Method Based on Serial QFD Neural Network

作　　者：赵师[1] 屈洋[1] 刘洪坤[1] 

机构地区：[1]蚌埠坦克学院,安徽蚌埠233050

出　　处：《指挥控制与仿真》2009年第5期103-106,共4页Command Control & Simulation

摘　　要：针对传统优化方法不能对武器装备使用主体――部队的使用需求进行有效考虑,提出武器装备使用需求到性能指标映射的串联QFD神经网络方法。该方法综合利用了QFD方法采集信息准确快速、神经网络能够精确模拟多变量输入输出强非线性关系的优点,提高了映射精度与针对性,克服了传统需求分析侧重于推理而轻实践、侧重于"设计方"而轻"使用方"的弊端,集中体现了"用户"为主的思想。改进了传统BP网络训练算法,给出了武器装备使用需求映射适应度函数。最后,结合实例对方法及算法进行了可靠性验证。Aiming at traditional methods can＇t study the force who are the user main body of the weapons and equipment efficiently, it puts forward a method based on serial QFD neural network which can map application demand to performance index. The method synthesis utilize the merit of QFD picking information accurately and neural network simulating intensive non-linearity relationship of multivariable input and output, it increases the mapping accuracy as well as targeted of application demand, makes up the traditional demand analyzing method which pays special attention to ratiocination and designer rather than giving full play to the experience and customers, and embodies the thought of under customers. The method makes up the traditional BP network algorithm, and puts forward a mapping function of equipment application demand. At last, it verifies the feasibility of the method in consideration of practical application.

关 键 词：神经网络 武器装备 使用需求 QFD 

分 类 号：E92[军事—军事装备学]