Reinforcement Learning-based Adaptive Fuzzy Controller for Coupled-Tank Liquid System Optimized by Advanced Jaya Algorithm

Abstract

This paper proposes a new reinforcement learning-based Fuzzy controller optimized with an advanced Jaya algorithm used as a robust control approach applied in coupled-tank liquid system. First, the parameters of fuzzy controller are optimally identified using Jaya algorithm based on reinforcement learning. Second, an adaptive fuzzy-based sliding mode surface is innovatively designed to guarantee that the closed-loop system is asymptotically stable using Lyapunov stability principle. The proposed algorithm is applied to accurately and robustly control the liquid level of the coupled-tank system. The comparison results with inverse fuzzy controllers and adaptive fuzzy controllers are fully presented to demonstrate that the proposed controller ensures an efficient and robust approach to effectively control the highly nonlinear uncertain systems.