On 23 June, the abstract
“Predicting Crew Absences for Railway Operations Using a Hybrid CNN–LSTM Model”*
will be presented by Ricardo Saldanha, Optimisation Leader, at
TransitData 2026 – International Symposium on the Use of Public Transit Automated Data for Planning, Operations, and Management, taking place in Toronto, Canada, from 23 to 25 June.
The abstract presents a deep learning model to forecast crew absences that can help dispatchers perform more accurate capacity planning, aligned with operational needs.
More reliable forecasts enable:
-
better decisions,
-
helping reduce overtime,
-
improve resource allocation, and
-
minimising the risk of service disruptions.
Tested on real railway data, the approach outperforms alternative models and supports better decision-making for planners and dispatchers.
More information here.
* Authored by André Filipe Leitão, Miguel Salvado, Luis Albino, Rui Rodrigues and Ricardo Saldanha