Akkaranee Phukdeewongand and Kannapha Amaruchkul
Thai Journal of Operations Research. Vol 1, No 1 (2013)
In this case study, we determine the cash levels in automated teller machines (ATMs) operated by one of the major commercial banks in Thailand. Our problem was formulated as a newsvendor model, in which the objective was to maintain a service level of 98 percent. A random cash demand during one cycle (3.5 days) was constructed by two different methods: 1) based on the total amount of cash withdrawals ; and 2) modeled as a random sum of the amount of cash withdrawals per each transaction, compounded by the total number of transactions. We analyzed 24-month historical data from 11 ATMs continuously operating at various places. Our study reveals that the policy based on method 2 outperforms the policy based on method 1 and the policy that the bank currently implements. Specifically, the expected cash level and the expected cash leftover could be reduced by 35.66% and 59.01%respectively, if the policy based on method 1 were implemented. Thus, better cash flow
management and less capital tied up could be achieved.
Keywords: Stochastic model applications; Inventory model; Supply chain management