EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
3924. Machine Learning of Chance Constraints for Scheduling Surgeries
Invited abstract in session TB-10: Surgery Scheduling and Operating Room Planning, stream OR in Health Services (ORAHS).
Tuesday, 10:30-12:00Room: 11 (building: 116)
Authors (first author is the speaker)
| 1. | Thomas Adams
|
Abstract
We describe several methods for approximating chance constraints on the duration of surgeries assigned to a surgical session.
We use historical data to create an empirical distribution of the session duration and employ SVMs and Probit regression to learn the probability that a session runs overtime, given the surgeries scheduled in it. These models are chosen as linear constraints can easily be formed from the results.
Through numerical experiments, we demonstrate the accuracy of these models to act as binary classifiers, and their effectiveness as approximations to chance constraints. We compare them to a distributionally robust approach to approximating the chance constraints, and show that the they result in similar, in many cases exactly the same, schedules but require much less time to solve the optimisation problem.
Keywords
- Health Care
- Scheduling
- Stochastic Optimization
Status: accepted
Back to the list of papers