In the interest of assessing the performance of PS agents at standard reinforcement learning tasks, we have created an interface that allows them to integrate with the OpenAI Gym. In order to use this functionality, you must first install the python package gym (following instructions provided on the project’s homepage.) Once this is done, you can use env_openai.py like any other environment for PS simulations, specifying the name of any OpenAi Gym task environment as an argument.