Data-Driven Nonlinear Near-Optimal Regulation Based on Multi-Dimensional Taylor Network Dynamic Programming

Abstract

Using the data-driven control formulation, an iterative dynamic programming approach which is based on a multi-dimensional Taylor network is established to design the near optimal regulation of discrete-time nonlinear systems. For discrete-time general nonlinear systems, the iterative adaptive dynamic programming algorithm is developed and proved to guarantee the property of convergence and optimality. Three networks are constructed, namely, the identification network, critic network and action network. Moreover, a globalized dual heuristic programming technique with detailed implementation is developed. The cost function and its derivative can be approximated by this novel architecture. Besides, without the consideration of the system dynamics, this technique can learn the near-optimal control law simultaneously and adaptively. In addition, this technique greatly improves the existing results of the iterative adaptive dynamic programming algorithm in terms of reducing the requirement of the control matrix. Furthermore, because of the approach that is based on the multi-dimensional Taylor network, the amount of calculation needed is also greatly reduced. The simulation experiment is described to illustrate the effectiveness of the data-driven optimal regulation method proposed in this paper. INDEX TERMS Adaptive dynamic programming, data-driven control, multi-dimensional Taylor network, nonlinear control.

Proactive Content Caching Based on Actor–Critic Reinforcement Learning for Mobile Edge Networks

“…There are some well-known approaches to solve the MDP problems such as dynamic programming [29]. Unfortunately, these approaches depend on perfect knowledge of the dynamic variation of environment to a great extent.…”

Adaptive Dynamic Programming-Based Fault-Tolerant Position-Force Control of Constrained Reconfigurable Manipulators

“…This part introduces the online PI algorithm [43] [44] to solve the HJB equation. The online PI algorithm is composed of the policy evaluation (26) and the policy improvement (27).…”

“…As an appropriate application for solving the optimal control problems of nonlinear systems, ADP algorithm, which was first proposed by Werbos [21], was considered as an effective approach in avoiding the difficulties of the 'curse of dimensionality'. Today, the ADP-based methods are being utilized in the designing of continue-time [22] [23], discrete time [24] [25] [26], data driven-based [27] [28] intelligence systems, and the solution of nonlinear optimal control with input/output constraints [29] [30], external disturbances [31] [32] [33] and actuator failures [34] [35]. Several investigations studied, the optimal control problems of the robot manipulator systems based on the ADP approach.…”

Multidimensional Nonlinear Landscape Planning Based on Parameter Feature Extraction and Multimedia Technology

“…Life is full of digital signals and digital technology. As the main bad environment space of our life, urban landscape is inevitably a ected and impacted by digital technology with its high sensitivity to science and technology [2].…”