返航备降航班高风险频发子集搜索模型

返航备降航班是一种保证航空安全的有效措施,但返航备降航班会带来成本损失,所以对返航备降业务规律进行分析,从而重点监控,在确保安全生产的基础上,有效降低返航备降航班的经济损失。我们可以把确定返航备降航班发生的密集时间段描述为一个多属性空间搜索问题,在许多实践领域存在该类问题,比如商业银行对贷款客户的信用评级、绩效考核等。在多属性组成的多维空间中,存在许多正常点和不正常点,通过建立高风险频发子集搜索模型和算法确定不正常点比较密集的多属性范围,使得在该空间范围内不正常点的比例较高并且总的点的数量达到一定的水平。本文以航空公司安全生产为例,其模型结果在实际生产的运用中取得良好效果,在一年内避免30余班次返航备降。

The Search Model on the High Frequent Risk Subsets of the Diversion or Turning Back Flights

The diversion or turning back flight is an effect method to ensure the flight safety,but it would be costly,so we can analyze the diversion or turning back flights and determine the frequently happening time period,and enhance the control.Thus,we can reduce the cost caused by the diversion or turning back flights on the basis of flight safety.The determination of the frequently happening period of the diversion or turning back can be described as the search problem on the multi attributes space,and its problem exists in many application areas,suchas,the customer credit rating for the commercial bank and the performance assessment.In the multi attributes space,there are many normal point and abnormal points,and we can determine the attributes range where the abnormal points frequently happen by mathematics model and programming and make sure the proportion of the abnormal points is much higher and the number of the abnormal points can get a certain level.The model has been applied to one Airline Company and more than 30 diversion or turning back flights have been avoided in one year since the application.