基于KPI和神经网络的作业车间绩效评价  

Performance evaluation of job shop based on KPI and neural network

作　　者：殷生旺 张月霞[1] 戴佐俊 YIN Shengwang;ZHANG Yuexia;DAI Zuojun(School of Information and Communication Engineering, Beijing Information Science ＆ Technology University ,Beijing 100101 ,China;The Company of ChuangYuan Electronics Suzhou 215299,China)

机构地区：[1]北京信息科技大学信息与通信工程学院,北京100101 [2]江苏创源电子有限公司,苏州215299

出　　处：《北京信息科技大学学报（自然科学版）》2018年第3期10-14,共5页Journal of Beijing Information Science and Technology University

基　　金：国家自然科学基金资助项目(51334003;61473039);北京市属高等学校高层次人才引进与培养计划项目(CIT&TCD201504058)

摘　　要：对作业车间绩效评价问题,提出了一种基于KPI和神经网络的的绩效评价体系;建立了作业车间绩效评价体系(JOB-KPI),在此体系的基础上建立了基于BP神经网络算法绩效评价体系模型。该神经网络模型采用logsig函数为传递函数,引入了动量因子,使得权值修正具有一定的惯性,加快了整个网络的收敛速度。仿真结果表明,绩效评价模型能够对离散车间绩效做出有效的评估和预测,可以为企业的决策者提供参考。To evaluate the performance of job shop,a performance evaluation system based on KPI and neural network is proposed. The performance evaluation system of job shop（ JOB-KPI） has been established. On the basis of this system,a performance evaluation system model based on BP neural network algorithm is established. The neural network model uses the log-sig function as the transfer function and introduces the momentum factor,which gives the weight correction a certain inertia and accelerates the convergence speed of the whole network. The simulation results show that the performance evaluation model can effectively evaluate and predict the performance of discrete workshop,which can provide a reasonable reference for decision-makers.

关 键 词：绩效评价 BP神经网络 logsig函数 动量因子 权值修正 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]