基于粒子群算法优化BP神经网络的产品质量预测分析  

Forecast Analysis of Product Quality Based on BP Neural Network with Particle Swarm Optimization

作　　者：王超英[1] 钟辉 

机构地区：[1]东莞职业技术学院计算机工程系,广东东莞523808 [2]广东汇兴精工智造股份有限公司,广东东莞523000

出　　处：《计算技术与自动化》2017年第3期92-95,共4页Computing Technology and Automation

基　　金：国家自然科学基金资助项目(61573156);东莞职业技术学院示范建设专项资金资助(政201717)

摘　　要：产品制造的过程中存在较大的不确定性,从事前预测的角度出发,提出了一种结合现有产品合格率、合格率变化规律等相关数据,借助BP神经网络等数学建模思想,并加入粒子群算法加以改进的产品质量预测模型,此种预测模型与传统BP神经网络相比,通过与粒子群算法的优化结合,进一步提高了预测精度。Owing to exist larger uncertainty in the process of product manufacturing, this paper proposed an improved prediction model of product quality joined particle swarm optimization using mathematical modeling thought, such as BP neu- ral network, which is combined with existing product percent of pass, the qualified rate of change law and other relevant data. Compared with traditional BP neural network, new prediction model further improves the prediction accuracy through the combination with particle swarm optimization algorithm.

关 键 词：BP神经网络 改进 粒子群算法 产品质量预测 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]