How to control electric autonomous taxi fleets in an energy efficient way

How can the development of new technologies impact the future of personal urban mobility? What direction can we give to the advancements in fields such as automation and electric vehicles, so that they can become our allies in the development of a more sustainable world? This thesis tries to provide an answer to those questions by investigating the emergent world of AMoD. Such new technology promises to revolution our cities and the way we travel across them, by releasing fleets of autonomous electric taxis that are able to pick up passengers in few minutes upon request and carry them to wherever they might ask. Many are the market players as well as the academic researchers currently taking part in this new technological challenge, and if some of its aspects have already been addressed and studied in detail, there are still numerous open issues that need to be investigated. In order to make this technology interesting enough to be adopted, so that it can cooperate with the already existing forms of transport and hopefully, one day, even replace some of them, it is necessary to enhance it and optimize its functioning. In this research, a new algorithm for an efficient control of such fleets of autonomous electric taxis is developed and tested. It includes the optimal assignment of passengers to vehicles, a periodic redistribution of free taxis along the network operated in a predictive framework, and a smart handling of vehicles' charging. The policy developed here shows good results in terms of a reduction of mean customers' waiting time and distance travelled by the taxis when not transporting passengers, resulting in an increased quality of the service and in a reduced energy consumption of the overall system. Some of the questions that are raised on the subject still remain open after this research, thus constituting interesting starting points for possible future studies.