摘要
提出了面向过程行业的遗传算法优化的BP神经网络质量管理模型,在单输出人工神经网络模型应用广泛的背景下,利用多输出神经网络模型解决了多目标整体优化的人工神经网络的工业应用问题。以建筑材料混凝土生产的质量管理为例,选取塌落度和28 d抗压强度为优化目标,计算结果表明平均相对误差在4%以内,证明了此方法的可行性,为企业的质量管理提供了强有力的数据支持。
A quality management model of backpropagation neural networks improved by genetic algorithm is proposed. With the wide application of single-output artificial neural networks model,a multi-output model of multi-target optimization is proposed. The quality management of commercial concrete is taken as an example,where slump and 28 d strength compressive is selected as optimization target. The mean absolute error of the model is under 4%,which approved the feasibility of the model and provide the strong data support for the quality management of process industry enterprises.
出处
《山东化工》
CAS
2017年第23期160-162,共3页
Shandong Chemical Industry
基金
国家自然科学基金资助项目(No.51473102
21776183)
关键词
质量管理
智能制造
人工神经网络
多输出模型
quality management
intelligent manufacturing
artificial neural networks
multi-output model
