A Reinforcement Learning Approach to Solving Incomplete Market Models with Aggregate Uncertainty