General Workflow and Seperation of Jobs

[KevinBian107]

We want to achieve a few things for this cooperative/competitive MARL repo (the priorities are listed as well), came up these ideas with Andrew today at Starbucks (CR Andrew):

• Baseline Models, i.e. PPO, SAC (ensuring that they learn) #2
• Reward Engineering (see what are needed changes for learning) #4
• Adapting Environment to Team playing #8
• Different Algorithm Fight Against Each Other #6
• Cycle & Curriculum Learning #7
• Transfer learning (can we tarin one agent in single player and then adapt to team, playing?) #33
• MCTS (use alpha zero style look ahead learning) #5