2dilc-rl

KO | EN

gitlite — search

#javascript #python #react #llm #machine-learning #deep-learning #linux #docker #macos #windows #cli #awesome-list #security #open-source

(★ 11)

2D Iterative Learning Control with Deep Reinforcement Learning Compensation for the Non-repetitive Batch Processes

2dilc-rl 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# 2dilc-rl: 2D_Iterative_Learning_Control_with_Deep_Reinforcement_Learning_Compensation_for_the_Non-repetitive_Uncertainty_Batch_Processes

## Catalog
* env_sys: Two different batch environment--injection molding process and nonlinear batch reactor. 
* ILC_Controller:  Solving the control law for the two-dimensional iterative learning controller.
* DRL_Compensator: Training the two-dimensional DRL compensator.
* Trained_2D_ILC_RL_Controller: the designed and trained 2D ILC-RL controller.
* Supplementary_Material: Supplementary materials for the Journal.
* comparison_method: The comparison control scheme: PI-based indirect-type ILC from paper 'PI based indirect-type iterative learning control for batch processes with time-varying uncertainties: A 2D FM model based approach' Journal of process control,2019
## Getting Started
* Clone this repo: `git clone https://github.com/CrazyThomasLiu/2dilc-rl`
* Create a python virtual environment and activate. `conda create -n 2dilc-rl python=3.8` and `conda activate 2dilc-rl`
* Install dependenices. `cd 2dilc_rl`, `pip install -r requirement.txt`

## Calculation of the 2D ICL controller
Run the following command to obtain the control law of the 2D ILC Controller.
```
cd ILC_Controller
python controllaw_injection_molding_process.py or controllaw_nonlinear_batch_reactor.py 
```

## Training of the 2D DRL compensator
Run the following command to train the 2D DRL Compensator.
```
cd DRL_Compensator
python demo_nominal_injection_molding_process.py or demo_nominal_nonlinear_batch_reactor.py 
python demo_practical_injection_molding_process.py or demo_nominal_nonlinear_batch_reactor.py 
```
The training usually takes 4 hours for the injection_molding_process and 6 hours for the nonlinear batch reactor.

## Test for the control performance simulation 
Run the following command to test the final control performance.
* Injection Molding Process

```
cd Trained_2D_ILC_RL_Controller/Injection_Molding_Process
python demo_injection_molding_process.py
```
![image](https://github.com/CrazyThomasLiu/2dilc-rl/raw/master/Trained_2D_ILC_RL_Controller/Injection_Molding_Process/Injection_molding_output.jpg)

* Nonlinear Batch Reactor

```
cd Trained_2D_ILC_RL_Controller/Nonlinear_Batch_Reactor
python demo_nonlinear_batch_reactor.py
```

![image](https://github.com/CrazyThomasLiu/2dilc-rl/raw/master/Trained_2D_ILC_RL_Controller/Nonlinear_Batch_Reactor/Nonlinear_batch_reactor_output.jpg)

# Citation:
The 2D Iterative Learning Control with Deep Reinforcement Learning Compensation for the Non-repetitive Uncertainty Batch Processes was published in the Journal of Process Control.

```
Liu J, Zhou Z, Hong W, et al. Two-dimensional iterative learning control with deep reinforcement learning compensation for the non-repetitive uncertain batch processes[J]. Journal of Process Control, 2023, 131: 103106.
```

2dilc-rl

2dilc-rl 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

2dilc-rl

2dilc-rl 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README