Kletzander, L. (2026). Parallel Constraint Programming and Simulated Annealing to solve the Integrated Healthcare Timetabling Problem. Operations Research, Data Analytics and Logistics, 47, Article 200512. https://doi.org/10.1016/j.ordal.2026.200512
When complex problems like the Integrated Healthcare Timetabling Problem (IHTP) combine several sub-problems and objectives, they create a challenging environment for optimization methods. Parts of the problem like surgical case planning and patient admission pose several constraints on feasibility, while the combination of objectives opens a large optimization landscape. This paper proposes to combine different technologies that specialize in these different requirements. Constraint Programming (CP) is used to quickly provide feasible solutions and further optimize the assignment of optional patients, while Simulated Annealing (SA) is used to improve feasible solutions with multiple neighbourhoods and efficient evaluation. Mild parallelization in four threads allows to run these methods in parallel, restarting SA from increasingly better CP solutions. Different solvers are evaluated on different versions of the CP model, SA is tuned by an automated parameter tuner, and ablations of different components are analysed. The results show a robust method that easily finds feasible solutions and provides good quality in the short time given for optimization.
