The goal of this work is to define the theoretical basis that can help extend the PEACH technology for museum visits, based on a Markov decision process approach. In order to promote interaction among visitors coming to a museum in a group, a decentralized recommender system is needed. The research is aimed at studying decision-theoretic planning for groups under time constraints with possibly conflicting goals. Algorithms are designed that will be able to recommend actions to a group of people visiting a museum taking into account time constraints imposed on the group, different priorities of the visitors and uncertainty in the environment.