摘要
自适应动态规划(Adaptive dynamic programming,ADP)方法可以解决传统动态规划中的"维数灾"问题,已经成为控制理论和计算智能领域最新的研究热点.ADP方法采用函数近似结构来估计系统性能指标函数,然后依据最优性原理来获得近优的控制策略.ADP是一种具有学习和优化能力的智能控制方法,在求解复杂非线性系统的最优控制问题中具有极大的潜力.本文对ADP的理论研究、算法实现、相关应用等方面进行了全面的梳理,涵盖了最新的研究进展,并对ADP的未来发展趋势进行了分析和展望.
Adaptive dynamic programming （ADP） method can solve the problem of ＂curse of dimensionality＂ in the traditional dynamic programming, and has recently become a hot topic in the field of control theory and computational intelligence. For ADP method, a function approximation structure is used to estimate the performance index function, and then the approximate optimal control policy can be obtained based on the principle of optimality. As a kind of intelligent control methods with learning and optimization capabilities, ADP has great potential in solving the optimal control problem of complex nonlinear systems. This paper presents a comprehensive survey on the theoretical research, algorithm development, and related applications of ADP, which covers the latest research progress. It also analyzes and predicts the future development trend of ADP.
出处
《自动化学报》
EI
CSCD
北大核心
2013年第11期1858-1870,共13页
Acta Automatica Sinica
基金
国家自然科学基金(61034002
61233001
61273140)资助~~
关键词
自适应动态规划
近似动态规划
强化学习
神经网络
智能控制
Adaptive dynamic programming （ADP）, approximate dynamic programming, reinforcement learning, neuralnetworks, intelligent control
