Randomized trial develops a decision support system for predictive maintenance in an industrial context, enhancing operational accuracy.
Key Points
This research aims to create an integrated system that combines multicriteria decision-making methods and machine learning techniques to improve asset maintenance decisions.
Conducted a structured case study with qualitative and quantitative data collection via interviews, focus groups, and observations.
Implemented multicriteria decision-making methods (AHP-AIP, MOORA, MULTIMOORA, Borda Count) and forecasting techniques (ARIMA, ANN) in a Python-based decision support system.
Prioritized assets for predictive maintenance and forecasted corrective maintenance needs.
Integration of MCDM and ML techniques improved decision-making accuracy for asset maintenance.
The developed decision support system is operational at the Brazilian Mint, effectively handling complex data challenges.
Fit analysis revealed the system delivers high value under varying conditions, demonstrating its potential adaptability beyond the specific organization.