ORAHS2025
Abstract Submission

143. A Two-Step Matheuristic for the Integrated Healthcare Timetabling Problem

Invited abstract in session TC-2: Integrated planning, stream Sessions.

Tuesday, 13:30-15:00
Room: NTNU, Realfagbygget R8

Authors (first author is the speaker)

1. Camille Pinçon
Mathematics and Industrial Engineering, Polytechnique Montréal
2. Nohaila Ahssinou
Polytechnique Montreal
3. Flore Caye
CIRRELT, Polytechnique Montréal
4. Prakash Gawas
Polytechnique Montreal

Abstract

The IHTC 2024 competition involves integrating three interconnected subproblems: Surgical Case Planning (SCP), Patient Room Assignment (PAS), and Nurse-to-Room Assignment (NRA). To tackle this highly complex problem, we adopt a two-step approach. First, we construct an initial feasible solution by sequentially solving each subproblem, ensuring both feasibility and computational efficiency. Next, we refine this solution using a Fix-and-Optimize (F&O) matheuristic, which iteratively enhances solution quality by selectively optimizing decision variables. The SCP phase focuses on scheduling feasibility while accounting for capacity constraints and workload balance. The PAS phase assigns patients to rooms based on gender and capacity constraints, employing two alternative strategies to maximize feasibility. The NRA phase ensures continuity of care and workload balance through a segmented planning horizon approach. Additionally, we apply a set of targeted destroyers that unfix parts of the feasible solution and then resolve specific subproblems to improve overall quality. Parallel computing is leveraged to accelerate solution generation and refinement. Our methodology is implemented using the JuMP library in Julia, with Gurobi as the optimization solver, running on multi-core architectures. This approach effectively balances feasibility, computational efficiency, and solution quality.

Keywords

Status: accepted

Back to the list of papers