Within companies, production is directly concerned with the manufacturing schedule, but other services like sales, maintenance, purchasing or workforce management should also have an influence on this schedule. These services often have together a hierarchical relationship, i.e. the leading function (most of the time sales or production) generates constraints defining the framework within which the other functions have to satisfy their own objectives. We show how the multi-agent paradigm, often used in scheduling for its ability to distribute decision-making, can also provide a framework for making several functions cooperate in the schedule performance. Production and maintenance have been chosen as an example: having common resources (the machines), their activities are actually often conflicting. We show how to use a fuzzy logic in order to model the temporal degrees of freedom of the two functions, and show that this approach may allow one to obtain a schedule that provides a better compromise between the satisfaction of the respective objectives of the two functions.