Multi-Objective Q-Learning

Reward Vector

Multidimensional reward vector where each component corresponds to the reward associated with a specific objective, replacing the traditional scalar reward signal.

← Indietro