RT - Journal Article T1 - Designing an Optimization Model for Inventory Control of ATMs in Tehran JF - JAMLU YR - 2020 JO - JAMLU VO - 17 IS - 2 UR - http://jamlu.liau.ac.ir/article-1-1684-en.html SP - 47 EP - 70 K1 - Data Mining K1 - K-Means Clustering K1 - Decision Tree K1 - Simulation K1 - Optimization K1 - Tabu Search. AB - ATMs are one of the most important cash-distribution channels, their inventory management is some of the most important tasks of banks. This paper provides a dynamic and optimized model for managing ATM inventory, based on the time and location of the devices. The number of devices examined is 368 in Tehran, which were surveyed during the quarterly period in 2018, which includes 189657 records. This model with data clustering succeeds in learning the pattern in the data, and then the proposed decision tree can predict the number of costumers for each scenario. By simulating the obtained scenarios, the system costs are determined, and finally, by optimizing the proposed model for each scenario, the average cost of the whole system is reduced by 12%. LA eng UL http://jamlu.liau.ac.ir/article-1-1684-en.html M3 ER -