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

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.