Multi-Agent Reinforcement Learning (MARL) - Independent PPO (IPPO)

[MaxGalindo150]

Objective:
This PR introduces the foundational components for the MARL module and implements the first multi-agent algorithm, Independent PPO (IPPO).

Key Changes:

• Added the MultiAgentRolloutBuffer for storing experiences across multiple agents.
• Implemented MultiAgentPolicy for decentralized policy management.
• Adapted the environment wrapper to support multi-agent scenarios.
• Developed the IPPO algorithm with unit tests and basic environment validation.

To-Do:

• Implement MultiAgentRolloutBuffer.
• Develop MultiAgentPolicy.
• Create a MultiAgentEnvWrapper.
• Implement IPPO algorithm.
• Validate IPPO in cooperative and competitive scenarios.
• Add basic documentation for MARL module.

---

Next Steps:
Once IPPO is fully integrated, subsequent PRs will focus on:

1. Implementing MAPPO with centralized training and decentralized execution.
2. Adding support for other MARL algorithms such as MADDPG and QMIX.