A Benders decomposition approach for planning home blood donations

Blood donations are crucial for the health system. We consider the problem of planning blood donation services, where the donors are reached at home. The scope is to minimize the penalty for the unserved donors, while guaranteeing that the available resources for implementing the service are not exceeded and that the appointment preferences of the donors are met. We present an offline model for this setting, where the produced solution must be robust with respect to the availability of the donors, which is not known in advance and is managed in a stochastic way using scenarios. A Benders decomposition approach to solve this model is developed. The proposed algorithm is tested on real-life instances coming from the Milan department of the Associazione Volontari Italiani Sangue (AVIS).