Tag
This paper introduces THV-UCB, an algorithm for multi-objective bandit problems with slate selection, and establishes gap-free and gap-dependent regret bounds for hypervolume regret.