| Title | Intelligent Adaptive Optimal Control: Algorithms and Stability |
| Speaker | Prof. Huaguang Zhang |
| Chair | Dr. Marco Wiering |
Abstract
Dynamic programming is a very useful tool in solving optimization and optimal control problems by employing the principle of optimality. However, it is often computationally untenable, if not impossible, to run true dynamic programming due to the backward numerical process required for its solutions, i.e., the well-known “curse of dimensionality”. Adaptive dynamic programming (ADP) is a promising method proposed for overcoming this difficulty, the idea of which is to approach dynamic programming solutions by using function approximation structures, such as neural networks or fuzzy basis functions, to approximate the cost function and control policy. In this talk, we will present the historic review and the state of the art of ADP. Specifically, starting from the motivation of introducing ADP, we will review the variances on the structure of ADP schemes and state the development on the ADP algorithms. An on-line iterative neural network optimal tracking controller is developed for the unknown nonaffine nonlinear systems, and a fuzzy optimal control scheme is established based on generalized fuzzy hyperbolic model (GFHM) for a class of affine nonlinear system. Furthermore, considering that many real systems are controlled by more than one controller or decision maker with each using an individual strategy, we further propose a new pair of mixed optimal policies to solve the nonlinear games when the saddle point does not exist. Meanwhile, for the situation that the saddle point exists, the existence conditions of the saddle point are avoided. Through numerical study, the feasibility and effectiveness of these methods are demonstrated.
Biography
Prof. Huaguang Zhang received the Ph.D. degree from Southeast University, Nanjing, China, in 1991. From 1992 to 1994, he did his postdoctoral research at Northeastern University, Shenyang, China. He has been with Northeastern University since 1991, and is currently as a Full Professor and Head of Electrical Automation Institute, Northeastern University. He has authored and coauthored over 200 journal and conference papers, five monographs and co-invented more than 20 patents. He has been serving as an Associate Editor of Automatica since 2008, an Associate Editor of IEEE Transactions on Neural Networks since 2009, an Associate Editor of IEEE Transactions on Fuzzy Systems since 2008, and an Associate Editor of IEEE Transactions on Systems, Man, and Cybernetics – Part B since 2007. He served as Chair or Co-Chair of more than 30 international conferences. He was the recipient of National Excellent Young Science Foundation Award in 2003 and Excellent Young Professor Award of China in 2002. He was nominated as the Changjiang Scholar by China Education Ministry in 2005, and was awarded National Excellent Postdoctoral Research Fellow Award in 2005, China. He received a second prize of State Technological Invention in 2007 and a second prize of State Scientific and Technological Progress in 2010 from the Chinese government, respectively. He has been serving as the Chairman of IEEE Shenyang subsection and the Chairman of IEEE Computational Intelligence Society Shenyang Chapter since 2010. He has also been a Chair of the Adaptive Dynamic Programming and Reinforcement Learning (ADPRL) Technical Committee of the IEEE Computational Intelligence Society since 2012. His current research interests include Adaptive Dynamic Programming, Fuzzy System Theory, Fuzzy Control, Neural Network-Based Control, Adaptive Control, Chaotic Control, and Complex Industry Process Automation.
