ORAHS2024
Abstract Submission

263. Operating room scheduling with surgeon assignment: a mixed integer linear programming and machine learning approach

Contributed abstract in session HB-5: Operating Room Scheduling /2, stream Regular talks.

Thursday, 11:00-12:30
Room: Room S6

Authors (first author is the speaker)

1. Alice Daldossi
Dipartimento di Informatica, Università degli Studi di Torino
2. Roberto Aringhieri
Dipartimento di Informatica, Università degli Studi di Torino
3. Sara Cambiaghi
Matematica, Università di Pavia
4. Davide Duma
Dipartimento di Matematica "Felice Casorati", Università degli Studi di Pavia

Abstract

This talk introduces a multi-objective Mixed Integer Linear Programming (MILP) model to deal with surgical case assignment and sequencing scheduling problems. The former concerns selecting patients from a waiting list and assigning an OR block to each of them. The latter addresses the decisions about the order of surgeries to be executed in the same OR block and the assignment of surgeons based on their compatibility and expertise. Then, a metaheuristic is proposed to ensure near-optimal and timely solutions applicable in real-world hospital settings.
While the presented optimization model addresses the combinatorial complexity of OR scheduling, several machine learning (ML) techniques are used and compared to address the uncertainty of the surgery durations. Since the proposed approach leverages data-driven insights to deal with such uncertainty, a comparison among the ML methods is executed considering standard accuracy indices of literature, but also evaluating the impact of errors on the quality of the resulting MILP model’s solution.
Computational analyses on real instances of the Vascular Surgery department of the hospital Città della Salute e della Scienza in Turin are presented. First, the effectiveness of the proposed metaheuristic is proved by comparing the solutions provided by such ad hoc algorithms with those of a general-purpose solver. Then, the trade-off between the considered objectives is analysed to represent possible preferences of decision makers.

Keywords

Status: accepted


Back to the list of papers