This year I want to try and train a team of agents using reinforcement learning in python. I have some questions about controlling the 2D simulator server using a trainer. Using the trainer (offline couch) I can start a match with the command “(start)” and stopped it with “(change_mode time_over)”. However, I don’t know how one resets the game clock to 0, using the trainer, to create a new match?
Also is it possible to reduce the time each cycle takes? Shorter game durations would allow for more experience to be collected. Thank you.