摘要
强化学习是人工智能的重要分支,具备与环境直接交互学习能力并具有优化决策的特点。强化学习问世以来便受到科研学者广泛关注,并探索其与深度学习之间融合的可能性。强化学习应用广泛,渗透教育、医药卫生、制造业、金融等多个领域。通过总结值函数与策略更新两种深度强化学习算法的演变历程,分别探索强化学习在智能路径规划领域中的优化算法,探讨算法落地过程中的难点及发展方向。
Reinforcement learning is an important branch of artificial intelligence,which has the ability to directly interact with the environment to learn and optimize decision-making.Since its inception,reinforcement learning has been widely concerned by researchers and explored the possibility of integration with deep learning.Reinforcement learning has a wide range of applica⁃tions,infiltrating many fields such as education,medicine and health,manufacturing,and finance.By summarizing the evolution of the two deep reinforcement learning algorithms,value function and policy update,the optimization algorithms of reinforcement learning in the field of intelligent path planning are explored respectively,and the difficulties and development directions in the process of algorithm implementation are discussed.
作者
程浩鹏
朱涵
杨高奇
晏为民
王慧婷
Cheng Haopeng;Zhu Han;Yang Gaoqi;Yan Weimin;Wang Huiting(School of Computer Science,Civil Aviation Flight University of China,Guanghan 618300;School of Science,Civil Aviation Flight University of China,Guanghan 618300)
出处
《现代计算机》
2022年第21期1-10,共10页
Modern Computer
基金
中国民用航空飞行学院学生科研基金(XSB2022-069)。
关键词
强化学习
路径规划
深度神经网络
reinforcement learning
path planning
deep neural network