本代码基于以下论文：

Mathew Mithra Noel, B. Jaganatha Pandian, Control of a nonlinear liquid level system using a new artificial neural network based reinforcement learning approach, 

Applied Soft Computing, Volume 23, 2014, Pages 444-451, ISSN 1568-4946, https://doi.org/10.1016/j.asoc.2014.06.037. 

(http://www.sciencedirect.com/science/article/pii/S1568494614003111)

代码演示了复杂非线性系统的RL控制。

Code demonstrates RL control of a complex nonlinear system. 

利用不同系统的状态空间模型来代替相互作用的双罐液位系统的状态空间模型，并且可以通过调整该代码来控制其他非线性系统。

The state space model of the interacting two-tank liquid level system can be replaced by the state space model of a different system and the same code can be tuned to control other nonlinear systems.

大多数工业过程都具有固有的非线性特征。

Most industrial processes exhibit inherent nonlinear characteristics. 

因此，采用线性化模型的经典控制策略不能有效地实现最优控制。

Hence, classical control strategies which use linearized models are not effective in achieving optimal control. 

本文提出了一种基于人工神经网络（ANN）的强化学习（RL）控制非线性相互作用液位系统的策略。

In this paper an Artificial Neural Network (ANN) based reinforcement learning (RL) strategy is proposed for controlling a nonlinear interacting liquid level system. 

这种ANN-RL控制策略利用了人工神经网络的泛化、抗噪性和函数逼近能力以及最优决策能力。

This ANN-RL control strategy takes advantage of the generalization, noise immunity and function approximation capabilities of the ANN and optimal decision making capabilities of the RL approach. 

提出了两种求解一般非线性控制问题的人工神经网络（ANN-RL）方法，并将其应用于两种基准非线性液位控制问题，对其性能进行了评价。

Two different ANN-RL approaches for solving a generic nonlinear control problem are proposed and their performances are evaluated by applying them to two benchmark nonlinear liquid level control problems. 

对基于离散化状态空间的纯RL控制策略进行了比较。

Comparison of the ANN-RL approach is also made to a discretized state space based pure RL control strategy. 

对基准非线性液位控制问题的性能比较表明，ANN-RL方法可以获得更好的控制效果，这一点可以通过减少振荡、抑制干扰和过冲来证明。

Performance comparison on the benchmark nonlinear liquid level control problems indicate that the ANN-RL approach results in better control as evidenced by less oscillations,disturbance rejection and overshoot.

完整源码下载地址：

http://page2.dfpan.com/fs/4lcje2216291368a4b5/

更多精彩文章请关注微信号：