Forecast-Driven Two-Stage Stochastic Programming for Nurse Rostering in Mental Healthcare
Mental Health
Healthcare Operations
Sequential Decision Analytics
Python
Time Series Forecasting
Machine Learning
Quarterly Forecasting Forum
Date: Mar 13 2026 12:00-17:00
Event: Quarterly Forecasting Forum
Location: Cardiff Business School, Cardiff University, UK
Context
At the Quarterly Forecasting Forum (QFF), I presented our research on “Forecast-Driven Two-Stage Stochastic Programming for Nurse Rostering.” This presentation focused on the application of a proposed forecasting approach, Autoregressive Markov-switching regression (AR-MSR), to forecast occupancy levels in mental hospital wards. The simulated forecasts were then integrated into a two-stage stochastic programming model to optimize nurse rostering decisions, aiming to improve patient care and operational efficiency.