Improving healthcare: integrating potential task splits into home healthcare routing and scheduling.

This thesis explores the potential of the currently neglected aspect of task-splitting for home healthcare routing and scheduling, focusing on the design of routes and time-tables for caregivers providing services at client’s homes. Motivated by contacts in the field, a new problem variant is proposed, which allows for the split of (long) patient visits into two separate visits. Those split parts might have less strict visit requirements, related to a wider visiting interval, reduced duration,
or less restrictive caregiver requirements, resulting in an increased number of planning options for home healthcare companies. However, the introduced temporal restrictions between the split parts pose a challenge.
To address this problem, a time-indexed mixed-integer linear program formulation is developed. A computational study is conducted, comparing scenarios with and without task-splitting across a wide range of generated benchmark problems. The results demonstrate the high potential of task-splitting for home healthcare providers. It can help to obtain feasible plannings in case of staff shortages, and leads to significant reductions in operational costs. Additionally, it enables
caregivers to spend more time on direct patient contacts. Furthermore, factors contributing to the benefits of task-splitting are identified, such as wage differences between caregiver types and the tightness of visiting time-windows. However, the temporal constraints form the computational bottleneck for the developed time-indexed formulation, making the problem computationally challenging to solve for larger real-life sized instances.

Keywords: home healthcare routing and scheduling problem, task-splitting, synchronisation, mixed-integer linear programming, time-indexed formulation 1