[Modification] Convert discrete action space to continuous action space in MAPPO MLP Actor

[CarcozZo]

I’ve modified the existing MAPPO code from a discrete action space to a continuous action space, and enabling support for more complex environments via OpenAI Gym's Box space.
Please visit my repository, and feel free to open an issue: https://github.com/CarcozZo/MAPPO-continuous-action-space

---

我已将现有的 MAPPO 代码从离散动作空间修改为连续动作空间，并且通过OpenAI Gym的Box空间支持自定义环境搭建。
请访问我的仓库，欢迎提出issue：https://github.com/CarcozZo/MAPPO-continuous-action-space