This is a part of my Graduation Thesis. I devised a new Reinforcement Learning Algorithm that made RL stable and faster to converge by using Teleportation Mechanism and Reward Dissolving. This algorithm demonstrates the best performance. Apart from that, I attempted to obselete learning rate in RL, making RL easier to fine-tune.

友情提示

1. 如果遇到内存问题导致无法正常运行，请在高级系统设置中打开虚拟内存
2. 主程序是这些以算法名命名的python文件，整个项目都需要完整下载才能正常使用
3. 建议按照指定路径使用本项目