A Dynamic Programming Model For Cash Management In ATM Based On GA-SA And Its Empirical Study

In recent decades, the frequent appearance of liquidity problems within the banking system, including the liquidity crisis and excess liquidity, have plagued the relevant authorities and the academia continuously. As financial market condition and policy changing, people have gradually deepened the understanding on the commercial banks’liquidity management. However, due to the volatility of financial markets intensified in recent years, the conversion between liquidity crisis and excess liquidity becomes rapidly. It brings an enormous challenge to the commercial banks’ liquidity management. Thus enhancing the commercial banks’liquidity management has a very important practical significance.With the booming of commercial bank self-service relying on electronic equipments such as automatic teller machines (ATM), the bank’s investment in such electronic devices has increased gradually, and the cash provision management in the corresponding equipments turns to an important part in the liquidity management of commercial banks. However, there are rare scholars who give adequate attentions to this area in China. In this paper, we establish a stochastic programming optimization model for the ATM cash management using the stochastic programming optimal technology and considering the uncertain future economic factors based on China’s real economic situation.The paper uses VAR and ARMA to generate scenarios. Further, it employs the Genetic Algorithm and simulated Annealing Algorithm (GA-SA) to solve the model, and obtains the optimal solution finally. The empirical results show that the results of the optimization model are better than the bank’s real performance. Empirical studies have shown that the model has certain reference significance and practical value.