RAPCG-MetaRL

Resource-Aware Procedural Content Generation using Meta Reinforcement Learning for Real-time Game Environments

Overview

RAPCG-MetaRL integrates real-time hardware telemetry into a reinforcement learning reward signal, creating a feedback loop that teaches PCG agents to balance content quality with computational efficiency. The framework targets heterogeneous gaming platforms — from budget laptops to high-end workstations — without requiring separate builds.

---

Implementation Status

Component	Status
PPO/A2C Training Pipeline	✅ Implemented
Resource-Aware Reward Shaping	✅ Implemented
Hardware Telemetry (psutil/pynvml)	✅ Implemented
Solvability Optimization	✅ Implemented
MAML Meta-RL Controller	✅ Implemented
Adaptive Batch Scheduling	🔄 Proposed
Hybrid PCG Ensemble	🔄 Proposed
Unity/Unreal Integration	🔄 Proposed

---

Dependencies

Core (Required)

• Python 3.10
• PyTorch 2.1+
• stable-baselines3
• gym
• numpy, pandas, psutil, pillow

Optional

• nvidia-ml-py3 — GPU monitoring
• jupyter — Notebooks
• matplotlib — Figure generation

See requirements.txt for full list.

---

Citation

If you use this framework, please cite:

@repo{RAPCG-MetaRL,
  title={Resource-Aware Procedural Content Generation via Meta-Reinforcement
         Learning for Heterogeneous Gaming Platforms},
  author={Redwan Rahman},
  link={https://github.com/Red1-Rahman/RAPCG-MetaRL}
}

Please also cite the foundational work this project builds upon:

@inproceedings{khalifa2020pcgrl,
  title={PCGRL: Procedural Content Generation via Reinforcement Learning},
  author={Khalifa, Ahmed and Bontrager, Philip and Earle, Sam and Togelius, Julian},
  booktitle={Artificial Intelligence and Interactive Digital Entertainment},
  volume={16}, number={1}, pages={95--101},
  year={2020}, organization={AAAI}
}

---

Acknowledgments

• gym-pcgrl - Ahmed Khalifa et al.'s foundational PCGRL framework
• stable-baselines3 - High-quality RL implementations
• PCG Benchmark - PCG evaluation testbed

Contact

Redwan Rahman — rahman22205101127@diu.edu.bd
Department of Computer Science and Engineering, Daffodil International University

Code: https://github.com/Red1-Rahman/RAPCG-MetaRL

---

RAPCG-MetaRL — Resource-Aware PCG that adapts to your hardware. 🎮