심층 강화학습을 활용한 산업디자인

Abstract

Deep Reinforcement Learning has developed rapidly in recent years and is likely to be applicable to a variety of real-world problems. In this paper, we examine the problems of industrial design and construct a reinforcement learning environment. In addition, we apply a representative Deep Reinforcement Learning methodology to examine experimentally whether industrial design is possible without human intervention. In the experiments, we achieved increased rewards in the reinforcement learning environments and better design results for the given tasks.