114. A kernel based approach for the estimate-then-optimize surgery scheduling problem
Invited abstract in session FA-2: Surgery scheduling 2, stream Sessions.
Friday, 9:00-10:30Room: NTNU, Realfagbygget R8
Authors (first author is the speaker)
| 1. | Ricardo Otero
|
| KU Leuven | |
| 2. | Erik Demeulemeester
|
| KBI, KU Leuven |
Abstract
Efficient and robust surgery scheduling is crucial for optimizing hospital resources and patient flow. However, creating adequate schedules is challenged by the inherent uncertainties in surgery durations. This research addresses this problem using a kernel-based approach to estimate surgery duration behavior and then implementing a distributionally robust optimization (DRO) model to generate solutions robust to model mispesification. Our methodology integrates a residual-based DRO framework, which constructs ambiguity sets based on historical data, with kernel methods that capture complex non-linear relationships between contextual features (e.g., patient demographics, procedure type) and uncertainty in surgery durations. By defining data-driven similarity measures through kernel functions, we create ambiguity sets tailored to the specific context of scheduling decisions, leading to more accurate robustness. The resulting optimization model is reformulated as a mixed integer model that standard solvers can solve. Our approach is evaluated against traditional methods using computational experiments with real-world data, yielding better out-of-sample solutions.
Keywords
- Operating room planning and scheduling
- Analytics
- Healthcare logistics
Status: accepted
Back to the list of papers