The present invention relates to a control device (110, 502) for controlling a drive train (100, 504) of a hybrid vehicle on a route to achieve a control goal, comprising :
a rule-based selection device (514), one of the rule-based selection device (514) depending on the Vehicle status-configurable mode enabler (512), and an agent (280, 402, 510) in signaling connection with the mode enabler (512), the agent (280, 402, 510) having a configuration based on reinforcement learning, which is based on data recorded during operation or during a simulation of the hybrid vehicle, and the agent (280,402,510) is set up to select an operating mode from at least one fixed operating mode and/or at least one variable operating mode based on the configuration of the mode enabler (512). to make adjustments to the drive train (100, 504).