This article deals with the deployment and redeployment of emergency vehicles, which is one of the main problems in Emergency Medical Services (EMS) management. Although the deployment of emergency vehicles has been extensively addressed in the literature, the originality of this article is that it explicitly considers the redeployment of vehicles - in our case, ambulances - based on cyclic changes in service demand (e.g., based on population movements during the day from residential areas to the workplace and vice-versa). We introduce a multi-period approach to take into account such service demand changes. The vehicle deployment plan is chosen to simultaneously maximize coverage and minimize vehicle relocation between periods. We propose a mathematical model that allows small to medium instances of the problem to be solved. In order to tackle large instances corresponding to real-life situations, we propose two heuristic approaches. We then evaluate the performance of our heuristics using a set of randomly-generated instances.--Résumé de l'éditeur.