Abstract
The integrated healthcare timetabling competition 2024 is about developing algorithms for efficiently finding good solutions to a certain timetabling problem using limited computational resources. The problem is about deciding when to admit patients to a hospital, where to carry out their surgery, where to host them and which nurses to assign to their rooms. It is entirely deterministic but involves several hard constraints (such as having patients of different gender not share a room) as well as eight different soft constraints by means of penalties, e.g., for delaying patients, nurse workload excess or having not too many different nurses care for a patient.
In the talk I will explain the approach of Team Twente, which is mainly a hierarchical approach that uses uses mixed-integer programming at different places, as well as constraint programming and simulated annealing. I will explain lessons that we learned and that eventually led to our approach.
This is joint work with the other team members Daniela Guericke, Asal Karimpour and Ieke Schrader from CHOIR/IEBIS as well as Rolf van der Hulst from our group.
