Pattern Generation Policies to Cope with Robustness in Home Care

We consider the Robust Home Care problem, where caregiver-to-patient assignment, scheduling of patient requests and caregiver routing must be taken jointly in a given planning horizon, and patient demand is subject to uncertainty. We propose four alternative policies to fix scheduling decisions and experiment their impact when used as a building block of a decomposition approach. Preliminary experiments on large size instances show that such policies allow to efficiently compute robust solutions of good quality in terms of balancing caregivers' workload and in terms of number of satisfied uncertain requests.