Experience Projects IISE OR Division Case Hackathon
IISE OR Division Case Hackathon
- Context
- IISE Operations Research Division
- Period
- 2025 – 2026
- Role
- Team leader
- Topics
- Operations Research, Workforce Scheduling, Bayesian Forecasting, Stochastic Optimization, Two-Stage Stochastic Modeling, Verification and Validation
- Tools
- Python, AMPL, Gurobi, ChatGPT (Codex), Excel
Our team built an app that turns historical nursing home demand data into forecast-driven staff schedules using Bayesian forecasting and stochastic optimization methods.
Problem
Nursing home managers have to build staff schedules before they know exactly how many residents they will have, what levels of care those residents will need, or when demand will spike. That makes scheduling both operationally difficult and financially important. Understaffing can lead to last-minute overtime or contract labor, while overstaffing is a significant avoidable cost. For the IISE OR Case Study Hackathon, our team focused on turning this broad customer need into something tangible: a tool that could help nursing managers create realistic long-term schedules from their own historical patient demand data.
Approach
We researched nursing home staffing practices, demand patterns, and operational constraints, then built two connected models. The first was a Bayesian forecasting model that used historical patient demand to estimate future care needs. The second was a two-stage stochastic optimization model that translated those future demand scenarios into a recommended nursing schedule. The model chose stable weekly assignments for RNs, LPNs, and CNAs while accounting for the possibility of extra contract or overtime labor when actual demand exceeded the planned schedule.
Solution
We combined the forecasting and optimization pieces into an app prototype where a nursing manager could input previous patient demand data and receive a recommended schedule for the next planning period. The final tool produced actionable staff schedules while making the cost tradeoffs between regular staffing, overstaffing, and contract labor easier to understand.
Results
After performing well in the first phase of the competition, we were invited to present our final solution at the finals during the 2026 IISE Annual Conference in Arlington, Texas. Our team ultimately tied for 2nd place.