基于系统解耦的多阶段Q-学习最优控制方法、设备及介质
By employing a multi-stage Q-learning method for system decoupling, the optimal control of high-dimensional discrete-time systems is optimized, solving the problems of low solution efficiency and high complexity in traditional methods. This enables efficient and real-time control strategy decision-making, and is applicable to fields such as automated production and distributed energy management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIHANG UNIV
- Filing Date
- 2025-06-19
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies are inefficient and difficult to solve optimal control problems in high-dimensional discrete-time systems. Traditional methods are prone to getting trapped in local optima and have high computational complexity.
A multi-stage Q-learning optimal control method based on system decoupling is adopted. Through a non-cooperative game framework and dynamic decoupling calculation, multiple value functions are optimized, a feedback control strategy is designed, and the least squares method is used to iterate the approximate value function to realize the dynamic trade-off of the control strategy.
It significantly reduces computational complexity, improves the real-time performance and robustness of control decisions, and reduces computational resource consumption, making it suitable for complex control systems such as automated production, intelligent scheduling, and distributed energy management.
Smart Images
Figure CN120507989B_ABST