The optimization of infrastructure maintenance costs in Automated Teller Machine networks of financial institutions is a huge problem faced through business intelligence as artificial intelligence for decision-making. This paper addresses the issue to optimize costs through the application of systems with artificial intelligence. So, neural networks were applied to predict ATM failures based on the historical information of specified errors, number and amounts of transactions. Then forecasting was used to determinate failures, after, the optimal maintenance route that the technical personnel must travel was determined by means of a genetic algorithm. Finally, it was estimated that the reduction of maintenance costs when applying the proposed predictive maintenance methodology is around 200,000 USD for an ATM network of 500 devices of a financial institution in Ecuador.
|Title of host publication||6th IEEE Ecuador Technical Chapters Meeting, ETCM 2022|
|Editors||David Rivas Lalaleo, Monica Karel Huerta|
|Publisher||Institute of Electrical and Electronics Engineers Inc.|
|State||Published - 2022|
|Event||6th IEEE Ecuador Technical Chapters Meeting, ETCM 2022 - Quito, Ecuador|
Duration: 11 Oct 2022 → 14 Oct 2022
|Name||6th IEEE Ecuador Technical Chapters Meeting, ETCM 2022|
|Conference||6th IEEE Ecuador Technical Chapters Meeting, ETCM 2022|
|Period||11/10/22 → 14/10/22|
Bibliographical noteFunding Information:
This research was funded by projects CYTED 788 [REDTPI4.0-320RT0006], PLAGRI project by Telecommunications and Telematics Research Group (GITEL) from Uni-versidad Politécnica Salesiana, Cuenca, Ecuador.
© 2022 IEEE.
- Business Intelligence
- Genetic Algorithm
- Neural Networks
- Predictive Maintenance