New mobility services using autonomous driving technology are being considered for the purpose of solving social issues and creating new businesses related to logistics and mobility. To implement new services and make them sustainable, not only safety but also economy and convenience must be considered at the same time. However, these performance indicators can often conflict with each other, and it is difficult to find service specifications that satisfy all of them. We explore a mobility service project which is being run for a large corporation in Japan.  In this session we introduce a method to reproduce/ evaluate new mobility services such as on-demand taxi or ride-sharing service in a virtual environment and present an example of service optimization based on simultaneous evaluation of safety, economy, and convenience.