Genetic Algorithm with PID Control system

1. Control System I used

1.1 Control System

1.3 The Input of The Control System

• 4 periods of standard rectangular signal
• Each period was cut into 60 pieces (ticks)

1.2 The Output And PID Setting of The Control System

2. Execute Project

2.1 Environment Setup

pip install pipenv
pipenv shell --python 3.9
pipenv install

2.2 Run the Project

Open the src/main.ipynb file, and run all cells

3. Results

The result table is based on the following settings of genetic algorithm:

1. population number: 50

2. iteration number: 100

3. mutation probability: 0.1

4. crossover rate: 0.9

5. PID value boundary: [0, 1]

6. eta value boundary: [0.30001, 0.69999]

7. fitness function:

   

Control System Output	Chromosome Value	Control System Setting
	P / I / D / eta: 0.990 / 0.249 / 0.021 / 0.3001	initial_y_now / u_boundary: -2 / [-20, 20]
	P / I / D / eta: 1.000 / 0.301 / 0.025 / 0.3001	initial_y_now / u_boundary: 0 / [-20, 20]
	P / I / D / eta: 1.000 / 0.269 / 0.012 / 0.3001	initial_y_now / u_boundary: 2 / [-20, 20]
	P / I / D / eta: 0.938 / 0.249 / 0.030 / 0.3001	initial_y_now / u_boundary: -2 / [-50, 50]
	P / I / D / eta: 1.000 / 0.257 / 0.000 / 0.3001	initial_y_now / u_boundary: 0 / [-50, 50]
	P / I / D / eta: 1.000 / 0.251 / 0.006 / 0.3001	initial_y_now / u_boundary: 2 / [-50, 50]