[Problem] To calculate in advance usage fees for respective users in a service which is shared among the users and in which a meter rate charging scheme is employed. [Solution] The present invention receives, in addition to a departure place and a destination place, a boarding request including whether cooperative ride share is to be conducted for a taxi. According to the present invention, a fee precalculation system sets vehicle allocation while also allowing a new user to share a ride on a taxi which is already traveling. In addition, the probability of occurrence of further ride sharing while the taxi is boarded is obtained. Meanwhile, the fee precalculation system has incorporated therein an algorithm trained by means of machine learning regression by use of teacher data in which fare fees for users had been calculated subsequently for operation data regarding cases where a plurality of users had used a taxi. With respect to the boarding request, by use of trained machine learning regression, the fees to be paid by the users can be calculated in advance in consideration of the probability of ride sharing that may occur in the future.