-
Overview:
-
Battle system is one of the most vital parts of SaGa and costly to balance due to the huge volume of the game. We present a machine learning (ML) framework to help us balance the battle system of the latest title of the series under development, SaGa Emerald Beyond, with Reinforcement Learning (RL). We show how our game designers use the AI to help their daily work, the challenges, the detail of our in-house algorithm, and the engineering backend to deliver the system. Also, our system and algorithm are generally applicable to other games and address major key challenges on implementing reliable ML system within general production line. With this system, we can evaluate more than 700 stages in one day and find critical bugs, speeding up the QA process significantly.
Recommended
