ORAHS2024
Abstract Submission

206. Integrating Heuristics and Roommate Preferences into an OpenSource Framework for Patient-to-room Assignment

Contributed abstract in session FC-3: Patient and Resource Scheduling, stream Regular talks.

Friday, 14:00-15:30
Room: Room S2

Authors (first author is the speaker)

1. Felix Engelhardt
Research and Teaching Area Combinatorial Optimization, RWTH Aachen University
2. Tabea Brandt
Lehr- und Forschungsgebiet Kombinatorische Optimierung, RWTH Aachen
3. Christina Büsing
Lehr- und Forschungsgebiet Kombinatorische Optimierung, RWTH Aachen University
4. Daniel Müller
RWTH Aachen University, Research and Teaching Area Combinatorial Optimization

Abstract

We consider the dynamic management of a hospital ward with respect to the patient-to-room assignment (PRA). PRA is an operational problem with multiple objectives, including avoiding unnecessary workload, i.e. patient transfers, avoiding patient conflicts and respecting single room entitlements.

We focus on two extensions of the basic model: First, we propose variations of GREEDY heuristics that quickly compute initial feasible solutions. Second, we evaluate another possible objective, i.e. roommate preferences and showcase how these can be introduced into the existing modelling setup.

Both extensions are evaluated in a computational study that is based on real-world data provided by the RWTH Aachen University hospital. For the roommate preferences objective, we also provide some combinatorial results and consider the effect of different modelling decisions on the computational performance. Based on this, we derive recommendations for modelling multiple objectives in PRA.

The computational study is based on an OpenSource framework [1] for PRA based on Binary Integer Programming (BIP), implemented in Python and using Gurobi as a solver. The framework already provides options for solving both static and dynamic PRA, including transfers and single room entitlements as objectives.

[1] https://zenodo.org/records/10408092

Keywords

Status: accepted

Back to the list of papers