Hey, I don't know if Im late to the party but this is fascinating! I may buy Imperator for its modding capabilities alone.
That being said, I have some questions:
1) Let's say I want to train a learner to optimise ai_plan_goals values using some cost function during game time (meaning, checking every year or so if the score improved and changing the values accordingly along the gradient), would that be possible?
To clarify: The learner would sit outside the game and just optimise the script values based on the games parameters which update live.
2) Is there some sort of API to read game parameters (like score) live?
3) Are there actions not governed by the action planning system and will they be added later?
Thank you for the great work!
That being said, I have some questions:
1) Let's say I want to train a learner to optimise ai_plan_goals values using some cost function during game time (meaning, checking every year or so if the score improved and changing the values accordingly along the gradient), would that be possible?
To clarify: The learner would sit outside the game and just optimise the script values based on the games parameters which update live.
2) Is there some sort of API to read game parameters (like score) live?
3) Are there actions not governed by the action planning system and will they be added later?
Thank you for the great work!