Hybrid Heuristics for Multimodal Homecare Scheduling

We focus on hybrid solution methods for a large-scale realworld multimodal homecare scheduling (MHS) problem, where the objective is to find an optimal roster for nurses who travel in tours from patient to patient, using different modes of transport. In a first step, we generate a valid initial solution using Constraint Programming (CP). In a second step, we improve the solution using one of the following metaheuristic approaches: (1) variable neighborhood descent, (2) variable neighborhood search, (3) an evolutionary algorithm, (4) scatter search and (5) a simulated annealing hyper heuristic. Our evaluation, based on computational experiments, demonstrates how hybrid approaches are particularly strong in finding promising solutions for large real-world MHS problem instances.