Providing an optimization model of inventory control costs in Tehran ATMs

Since cost management is one of the most important tasks of organizations, cost management of inventory control system of ATMs is also one of the most basic tasks of banks. This article seeks to provide a dynamic and optimal model for controlling inventory costs of ATMs, according to the time and place of each device. Therefore, all data, related to the relevant bank in Tehran, which includes 368 ATMs, was used. Investigating the behavior of devices in the three-month period in 1396 has been done. This model has succeeded in learning the existing pattern in big data by clustering statistical data in place and time dimensions, and based on this, the proposed decision tree is able to predict the number of customers to each device. Then, using the cost function for the obtained scenarios, the system costs are determined. The total cost of the system includes the total hold cost of money, shortage cost and orderig cost  for each device. Finally, by providing an optimized inventory control model for each scenario, the total system costs are reduced by an average of 16.5 percent, or 38 million tomans per month.