Achieving Stable Training of Reinforcement Learning Agents in Bimodal Environments through Batch Learning

E. Hurwitz, N. Peace, G. Cevora

Bimodal, stochastic environments present a challenge to typical Reinforcement Learning problems. This problem is one that is surprisingly common in real world applications, being p…