二维非均匀信息扩散及其应用  

Two-Dimensional Uneven Information Diffusion and Its Application

作　　者：沈赟琦[1] 陆余楚[1] 

机构地区：[1]上海大学理学院,上海200444

出　　处：《应用数学与计算数学学报》2007年第2期91-96,共6页Communication on Applied Mathematics and Computation

基　　金：复旦一瑞士再保险研究基金

摘　　要：在风险分析和评估过程中,经常会遇到信息不充分、数据不完备的情况,即小样本问题.本文提出了二维的非均匀信息扩散方法,对小样本进行分析处理,并用于实际项目加以印证.其中非均匀信息扩散方程取(N)/(t)=(N+1)·((~2N)/(x^2)+(~2N)/(y^2)),用MacCormack方法进行求解,在两个优化准则下建立优化模型,利用MATLAB编程实现,结果与大样本是一致的.In the process of risk analysis and evaluation,the information is often scarce and the collectable data is uncompleted,which is called 'Small Sample Problem'. This paper analyzes and deals with two-dimension small sample problem by using Uneven Information Diffusion Method(UIDM).And UIDM is well verified by a practical project. We assume the uneven information diffusion equation be (N)/(t)=(N+1)·((~2N)/(x^2)+(~2N)/(y^2)),and make use of MacCormack method to get its solutions.Then,the optimization model is set up based on the two optimization principles associated with the mathematical software MATLAB.The results are coincidence with the large sample.

关 键 词：非均匀信息扩散 MacCormack方法 小样本 

分 类 号：O213[理学—概率论与数理统计]