基于粗神经网络的管制员工作差错风险预警模型  

Work error risks warning model of air traffic controller based on rough sets and neural networks

作　　者：杨智[1] 罗帆[1] 

机构地区：[1]武汉理工大学管理学院,武汉430070

出　　处：《中国安全生产科学技术》2011年第12期44-50,共7页Journal of Safety Science and Technology

基　　金：国家自然科学基金资助(编号:70971104)

摘　　要：针对空中交通管制员工作差错风险类型,指出指标体系建构过程的动态性特征,从个人、团队、设备、环境及管理几个方面,建构预警指标体系,运用粗糙集在数据挖掘方面的优势,通过基于属性重要度的启发式约简算法,提取关键指标,剔除冗余指标,可以实现在不丢失关键预警监控对象的情况下,结合BP人工神经网络构建实时预警模型,既有利于加快运算速度,又有利于进行重点监控。经过实例仿真,粗糙集与BP网络结合建构的预警模型,能有效针对管制员工作差错风险进行实时预警监控。According to the types of work mistakes risk for air traffic controllers,the dynamic characteristic in construction process of index system were pointed out.From the aspects of personal,team,equipment,environment and management,the early-warning index system of work mistakes risk was established.Using the rough sets in advantages of data mining technology,through the attribute importance of heuristic reduction algorithm,the key index was extracted,and the redundant index was eliminated.Combining with the BP artificial neural network,the real-time warning model was built in the case of without the loss of critical early-warning monitoring object.It not only can help to speed up the computational speed,but also help to focus on the key object monitoring.Through the sample simulation,combining rough sets with BP networks,the early warning model was constructed,which could effectively realize the function of real-time warning monitoring for work errors risk of air traffic controller.

关 键 词：管制员 工作差错 安全风险 粗糙集 BP神经网络 

分 类 号：X923[环境科学与工程—安全科学]