Abstract: Current multi-objective reinforcement learning (MORL) research often struggles to balance multiple objectives and manage the stability and performance of learning algorithms, especially in ...