While reinforcement learning algorithms converge towards a single policy, it may be useful to generate multiple policies instead of just one. read more of Diversity Algorithms