小批量生产过程中产品质量控制的研究  

Study on the Control of Product Quality in Small-scale Production Process

作　　者：张文生[1] 

机构地区：[1]中国科学院软件研究所

出　　处：《系统工程理论与实践》2002年第5期24-30,共7页Systems Engineering-Theory & Practice

基　　金：国家"8 6 3"高技术研究发展计划 ( 2 0 0 1 AA1 1 4 1 70 );国家自然科学基金 ( 7970 0 0 2 3;6 0 0 330 2 0 ) ;航空基础科学基金 ( 97J5 5 0 0 9)

摘　　要：借助支持向量机技术解决了下面三个问题 :1 .在给定核函数的前提下 ,客观地、系统地给出一组增强输入模式 ,进而构造指出向量机网络 ,求出小批量数据所遵循的规律 ;2 .通过理论分析说明 ,该方法不仅能控制学习结果对测试样本的误差 ,而且能提高学习结果的泛化能力 ,从而避免 BP学习过程的缺陷 ,;3.通过求解小批量生产过程产品质量规律的实例 ,验证了上述结果 。The rule of product quality is difficultly found by using test data in small-scale production process directly. This paper solves the three problems as below by means of SVM technology: 1. In the condition of specified kernel functions, objectively and systematically present a group of enhanced input mode, then construct the function-link network and give the result of the small-scale data; 2.Demonstrate from theoretical analysis that our approach not only controls the learning error, but also improves generalization ability of the learning result, and Avoid the deficiency of BP learning process ; 3. We verify our conclusion through seeking the rule of product quality in small-scale production process, and show that our approach can be used for real-time process greatly.

关 键 词：小批量生产过程 产品质量控制 函数型连接网络 支持向量机 现代制造业 

分 类 号：F407.4[经济管理—产业经济]