基于定性相空間的應(yīng)急資源需求預測方法研究.pdf

1、中國科學技術(shù)大學碩士學位論文基于定性相空間的應(yīng)急資源需求預測方法研究姓名：操張進申請學位級別：碩士專業(yè)：管理科學與工程指導教師：@2011-05-20Abstract III ABSTRACT Emergency logistics management, which has aroused growing concerns, has become a research hotspot in recent years due to it

2、s great significance for improving the emergency management capability of Government. After the unexpected events happened, emergency logistics distribution system not only needs to transport emergency resource, but also

3、 should timely supply appropriate goods to affected areas in right way. Because of hardness of timely acquiring disaster information and the imbalance between supply and demand, if decision-makers can initiatively integr

4、ate disaster information and predict the demand of emergency resource after the occurrence of large-scale disaster, it will contribute to allocate emergency resources effectively and efficiently and avoid the imbalance o

5、f supply-demand. In this paper, considering the characteristics of emergency demand time series, we proposed a simulation theory based algorithm which combines qualitative and quantitative models to forecast the demand a

6、fter unexpected events. This study selected the daily trading data from an agricultural products market after the 2008 snowstorm in Anhui province as the time series which will be studied in this paper. Firstly, this art

7、icle proved that the aforementioned time series is chaotic and applied the phase space based maximum Lyapunov exponent model to this time series to obtain preliminary forecast results. Then, we adjusted the results by us

8、ing qualitative knowledge. In addition, this paper applied fuzzy equivalent matrix based clustering analysis and TOPSIS to identify the damage level of each affected area, then we know the urgency of each disaster node a

9、nd this is necessary knowledge for emergency resource distribution. Finally, we concluded that the proposed prediction method in this paper performs better than ARIMA model and merely quantitative methods, and the propos