您的位置:首页 > 论文页面
基于K-means算法的GABP神经网络预测工序的工时定额
发表时间:2019-12-31 浏览量:1297 下载量:130
全部作者: | 樊梁华,雷琦 |
作者单位: | 重庆大学机械传动国家重点实验室 |
摘 要: | 为提升零件工时定额的准确性及实时性,本文考虑从零件的工序角度出发,提出一种K-means聚类结合遗传算法(genetic algorithm,GA)优化BP(GABP)神经网络预测工序工时定额的方法。首先根据工序工时的影响因素统计出历史及待预测的工序加工特征值;利用K-means聚类算法对历史的工序加工特征值进行聚类;对聚类分组的加工特征值分别建立对应的GABP神经网络预测模型并进行训练;根据欧氏距离的最小原则将待预测的工序加工特征值划分到对应的聚类组及预测模型;最后利用对应的模型预测待预测工序的工时定额。通过实例验证结合K-means聚类的GABP神经网络预测模型比未结合K-means聚类的GABP预测模型的准确率高,且预测误差控制在10%以内,证明了该方法的可行性与有效性。 |
关 键 词: | 计算机应用;工时定额;预测;工序;遗传算法优化BP(GABP)神经网络;K-means聚类 |
Title: | Working time quota of process predicted by GABP neural network based on K-means algorithm |
Author: | FAN Lianghua, LEI Qi |
Organization: | State Key Laboratory of Mechanical Transmissions, Chongqing University |
Abstract: | In order to improve the accuracy and real-time performance of working time quota of parts, a method of K-means clustering combined with BP optimized by genetic algorithm (GA) (GABP) neural network is considered to predict the process working time quota from the perspective of parts in this paper. Firstly, according to the influencing factors of process working time, the historical processing feature values and the processing feature values to be predicted are calculated. The K-means clustering algorithm is used to cluster the historical processing feature values. The corresponding GABP neural network prediction models are established for the clustered processing feature values. And the models are trained. According to the minimum principle of Euclidean distance, the process processing feature values to be predicted are divided into corresponding cluster groups and prediction models. Finally, the corresponding model is used to predict the work time quota of the process to be predicted. Verifying by the example, the GABP neural network prediction model combined with K-means clustering is more accurate than the GABP prediction model without K-means clustering. The prediction error is controlled within 10%, which proves the feasibility and effectiveness of the proposed method. |
Key words: | computer applications; working time quota; prediction; process; BP optimized by genetic algorithm (GABP) neural network; K-means clustering |
发表期数: | 2019年12月第6期 |
引用格式: | 樊梁华,雷琦. 基于K-means算法的GABP神经网络预测工序的工时定额[J]. 中国科技论文在线精品论文,2019,12(6):896-904. |

请您登录
暂无评论