Abstract: Reinforcement learning (RL) methods that train agents using simulation are well suited to solve behavioral decision-making problems. However, complex simulation platforms that have slow ...