fix controlling the n_parallel_agents and the concurrent env operations

[LianShuQuan]

Does n_parallel_agents means the number of active agents, and max_workers means the concurrent env operations?
If so, then
trajectory_generator() should use asyncio.Semaphore() like execute_tasks();
self.executor = ThreadPoolExecutor(max_workers=max_concurrency) should be modified in trajectory_generator();
self.n_parallel_agents = len(envs) should be deleted in update_envs_and_agents().

[jeffreysijuntan]

Great catch. How about renaming max_workers to max_env_workers then?

[LianShuQuan]

Great catch. How about renaming max_workers to max_env_workers then?

Sure. Should I only change max_workers to max_env_workers besides the initialization parameters in order to maintain compatibility with the initialization and usage code of AgentExecutionEngine?

[jeffreysijuntan]

Yes, simply rename this parameters to max_env_workers to make it clearer.