Q-learning with biased policy rules

Re-assign community