Reinforcement learning approach to solve a sequential decision making problem for designing optimal breast cancer screening policy