基于灰色马尔科夫模型的机场安检危险品数量预测

机场安检的危险品数量具有动态、随机、非线性等特点,传统的GM(1,1)模型无法对其作出准确的预测。利用灰色GM(1,1)模型对2014年1—5月所查获的危险品数量进行计算、检验,并对6—8月的危险品数量进行预测。首先建立危险品数量的GM(1,1)模型,然后再对其预测值进行修正,结果表明,灰色马尔科夫模型的平均相对误差比灰色预测模型的平均相对误差减小了25.18%,表明灰色马尔科夫模型比单一的灰色预测模型的精度高,该模型是有效可行的,可为航空公司6—8月将要查获的危险品数量预测提供理论基础,以便引起相关部门的高度重视,并采取相应措施以保障旅客安全。

Prediction of the amount of dangerous goods to be inspected by the airport security department via the grey Markov model

The purpose of this paper is to predict the amount of dangerous goods to be searched out and confiscated by the airport security inspection departments by using the GM (1,1) model.As is known,since the amount of dangerous goods to be detected and confiscated by the airport security department is dynamic,random in nature and non-linear in characteristics,it is impossible to be predicted in a traditional GM (1,1) model,which makes us predict and confiscate their amounts by using the gray prediction model based on the Markov gray model in this paper.In doing so,we have adopted the grey GM (1,1) model in calculating and verifying the amount of such dangerous goods that have been detected and confiscated by the airport security inspection departments in the first five months of 2014 as well as have made a prediction of the amount of such goods to be detected in the coming months of June to August.To do a good job in this way,we have first of all established and prepared a time response function model based on the theory of Markov model,and,then,we have determined the relative error between the actual amount and that for the predicted.The second step is to fix or frame its actual status-situ transition probability matrix through the size of the relative error between the predictive amount and that of the actually searched out,and then to identify and determine the actual gap between the two.Comparing the error gap between the actual detected amount and that of the predicted by using the grey GM (1,1) model,we have found that the relative gap can be reduced to 25.18％.Thus,the gap can help to show that the grey GM(1,1) model tends to be more accurate than the grey prediction model we have developed has done.The results also prove that the prediction model is feasible and qualified for its mission.For one thing,though the grey Markov model may seem to be more endurable from the long-term forecasting results for its takes in some advantages of the gray prediction model.For the other,it can provide more accurate predictive volatility data,due to its absolutely forecasting power.What is more,it can help to lay a theoretical basis in predicting the amount of the dangerous goods that have been searched out and fined by the airport security inspection departments from June to August,which tends to be in a rising tendency in size and amount.Therefore,the tendency has to be paid greater attention to by the departments concerned,and in turn,more proper measures have to be taken so as to reduce the amount of such risk-involved goods and increase the safety both for the society and the passengers.