基于遗传模拟退火算法的ATM备付金管理动态规划模型及实证研究

发布时间：2018-01-05 06:37

本文关键词：基于遗传模拟退火算法的ATM备付金管理动态规划模型及实证研究　出处：《东北大学》2012年硕士论文　论文类型：学位论文

【摘要】：近几十年来,银行系统内部频繁出现的流动性问题,包括流动性危机和流动性过剩问题,一直困扰着有关当局和学术界。随着金融市场环境和国家政策的改变,人们已逐步加深对商业银行流动性管理的理解,但是由于最近几年金融市场各项要素的波动性加剧,商业银行的流动性变动加快,流动性危机能够轻易和流动性过剩相互转换,给商业银行的流动性管理造成重重困难,进一步规范和强化商业银行流动性管理势在必行。随着依托于自动取款机(ATM)等电子设备的商业银行自助服务的蓬勃发展,银行对此类电子设备的投入逐渐加大,相应设备中的备付金管理也成为流动性管理中不可或缺的重要部分。但是,国内少有学者对该领域的研究给予应有的重视。本文基于随机规划优化技术,结合国内银行业的真实状况,考虑不确定的未来经济因素,在合理的约束条件下,建立了一个具有一般性的ATM备付金管理的动态规划优化模型。对于模型的求解,本文基于历史数据,通过向量自回归(VAR)、自回归滑动平均混合模型(ARMA)和线性方程等方法进行情景生成,使用遗传模拟退火算法来求解模型,并得出最优解。实证结果表明,优化模型的结果优于银行的真实表现,因此,本文所建立的模型具有一定的参考意义和实践价值。
[Abstract]:In recent years, frequent internal liquidity problems in the banking system, including the liquidity crisis and excess liquidity problems, has been plagued by the relevant authorities and the academic circles. As the financial market environment and the changing of national policy has been gradually deepen to the understanding of the liquidity management of commercial banks, but due to the volatility in recent years the elements of the financial market intensifies, the liquidity of commercial banks to speed up the changes, the liquidity crisis easily and excess liquidity conversion, liquidity management of commercial banks caused difficulties, further standardize and strengthen the liquidity management of commercial banks is imperative.
With relying on the automatic teller machine (ATM) rapid development of electronic devices such as commercial banking services, the bank put on such electronic equipment gradually increased, the corresponding equipment in the reserve management has become an important part of the liquidity management. However, few scholars in this field should be given pay attention to. This paper optimization technique based on stochastic optimization, combined with domestic banks in the real situation, considering the uncertainty of future economic factors, in reasonable constrains, established a dynamic programming optimization model of a general ATM reserve management.
For the solution of the model, based on the historical data, through vector auto regression (VAR), the autoregressive moving average model (ARMA) and linear equation method of scenario generation, using genetic simulated annealing algorithm to solve the model, and optimal solution. The empirical results show that the optimization model of real performance, better than the bank therefore, the model has certain reference significance and practical value.

【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F832.5;TP18

【参考文献】