A data-driven approach was developed for selecting resource allocation policies in a medical device reuse system (inspired by one of the use cases), using MDP modelling and machine learning to balance holding and shortage costs under capacity constraints. This work, presented at EURO 2025 in Leeds, demonstrates that AUTO-TWIN’s method enables efficient and automated policy selection based on problem parameters.

https://www.theorsociety.com/ORS/ORS/Events/2025/EURO-2025/EURO-2025.aspx