Iteratively Learn Diverse Strategies with State Distance Information