This is the open source code of paper "CMA: Collaborative Microservice Autoscaling Based on Reinforcement Learning in Edge-Cloud Environment"
Directory Structure: "algos" and "onpolicy" folders store algorithm-related files. runner is the entry point for functions. Within it, GMPERunner().train() is used to train the algorithm, and GMPERunner().eval_cma() is used to evaluate the algorithm. Potential Issues:
- Incorrect directory structure.
- Issues caused by path changes.