A method, scheduling server and system for improving automated scheduling, wherein the automated scheduling is at least partially based on user settings input by a plurality of users includes: obtaining user settings for a plurality of users, the settings defining user preferences and personal constraints of each user, obtaining system boundary conditions derived from a schedule goal and technical conditions of the technical equipment involved to achieve a desired goal that shall be achieved by applying the schedule to be determined and calculating a schedule by solving a scheduling problem specified from a combination of the user specific settings and the system boundary conditions. Then, acceptance probabilities for alternative user settings are predicted, which are settings differing from the obtained user settings. At least one alternative user setting is selected and an alternative schedule for each selected alternative setting is calculated. The alternative schedules are compared to the initially calculated schedule with respect to quality measure, and, based on the result of the evaluation and the predicted acceptance probabilities, an alternative setting to be suggested to the user whose setting is concerned by the alternative setting is determined. After reading in the user's response to the suggested alternative setting another alternative setting is selected in case the user declined, or the systems proceeds with accepting the alternative setting as obtained user specific setting and repeats calculation of a schedule, selection of alternative settings, calculation of alternative schedules, evaluation and select and suggest an alternative setting and read in of the users response until a stop criterion is met, and apply the last calculated schedule.