It is done to control Netlogo from python on the Ubuntu system. Netlogo is a helpful visual tool for multi-agent simulation. However, it is challenging to analyze further the agent behaviors and action policies. Python is a better choice to cover Netlogo. Furthermore, it can easily construct a reinforcement learning mechanism for policy enhancement. The point is to build an efficient shared environment for these two.
After many trials, I used Jpype and pyNetlogo in the latest Ubuntu system.

How to build pyNetlogo environment in Ubuntu

Install anaconda

To download the latest anaconda installation package.
https://www.anaconda.com/products/distribution#linux

ref: https://blog.eldernode.com/setup-anaconda-on-ubuntu-22-04/

$ cd Downloads
(change downloaded file name to a simple one.)
$ bash Anaconda3.sh 

To create a desktop shortcut for Anaconda Navigator
ref: https://www.how2shout.com/linux/create-anaconda-navigator-desktop-shortcut-ubuntu-20-04-18-04/

Generate pyNetlogo env in anaconda

Intsall numpy, scipy, pandas

In terminal, $ conda activate pyNetlogo

Install JPype

Because JPype is not in conda list, confirm it in the official site.

1	`$ conda install -c conda-forge jpype1`

note: JPype can not work in Python 3.9, but fine in 3.8.

Install pyNetlogo

ref: https://pynetlogo.readthedocs.io/en/latest/install.html

note: pyNetLogo requires NumPy, SciPy, Pandas and JPype packages.

$ pip install pynetlogo
Conda installs matplotlib and seaborn

note: do it later.change in core.py of pyNetlogo

Install Java JDK

ref: here

$ sudo apt update  
$ sudo apt install -y openjdk-17-jdk

Choose JDK17 to install.
The Java installation costs time.
After installation, check the Java version: $ java –version

End processes while error. $ gnome-system-monitor

Summary

Key steps of installation:

$ conda install -c conda-forge jpype1
$ pip install pynetlogo
$ sudo apt install -y openjdk-17-jdk

Notes:
(1) JPype does not work on Python 3.9, but fine on 3.8.
(2) Skip JDK19 trial while JDK17 works.