• 목록
• 게시판 리스트 옵션
  • 검색

In multistage manufacturing processes, quality can vary significantly depending on the chosen path, due to the difficulty of ensuring uniform product quality from each piece of equipment. We introduce the use of Monte Carlo Tree Search (MCTS) method to consider quality in addition to productivity in the path selection process. 

Specifically, we propose a method to incorporate both short-term and long-term benefits as rewards in the MCTS model by utilizing main effects and interaction effects from traditional experimental designs, to produce paths that maximize yield. This approach is expected to allow optimal scheduling while ensuring quality even in large-scale multistage process environments where conventional methods may struggle to estimate quality outcomes.