PSA CodeSprint 2025: Smart Port Operations – AI for Horizontal Transport Optimisation (Node)

Design an AI-driven scheduling and routing strategy to minimize congestion, balance yard utilization and improve overall efficiency.

Environment Setup

Before working on the problem, please create a Python virtual environment and install all required dependencies.

If you are familiar with Python environments, you can directly install from requirements.txt. Otherwise, follow these step-by-step instructions in your Command Prompt (Windows):

# Create a Python virtual environment. Here I used Python 3.11.
python -m pip install virtualenv
python -m virtualenv py_env_codesprint
py_env_codesprint\Scripts\activate

# Install pip-tools to manage dependencies
python -m pip install pip-tools

# Compile and synchronize dependencies from requirements.in to requirements.txt
pip-compile -v --rebuild -o requirements.txt --trusted-host pypi.org --trusted-host pypi.python.org --trusted-host files.pythonhosted.org
pip-sync requirements.txt

# (Optional) Set up Jupyter kernel for experimenting with notebooks
python -m ipykernel install --user --name=py_env_codesprint_kernel

To activate your environment:

py_env_codesprint\Scripts\activate

Development

To get started, run the simulation with a GUI by executing:

python gui.py

You will see a simulation of the port terminal operations.

About the Code Base

The simulation consists of two main components:

• Planning Engine: Where scheduling and navigating algorithms are implemented.
• Operation Engine: Simulates execution of planned jobs.

Your task is to improve the default planning algorithm in the Planning Engine for better performance.

Project Structure

INA_DS_CS/
│
├── data/
│ └── input.csv # Input job details for the planning algorithm
├── logs/       # Log files for each simulation run
├── src/        # Source code
│ └── operate/
│ ├── plan/
| |  └── job_planner.py # Your main implementation goes here!
│ ├── ui/
│ └── utils/
├── cli.py      # Command-line interface entry point
└── gui.py      # GUI interface entry point

• input.csv contains the job details fed into the planning algorithm.
• job_planner.py is the core of the Planning Engine — your modifications should be made here (Search for YOUR TASK HERE). Refer to the provided documentation for details on the default algorithm logic.
• gui.py runs the simulation with a graphical interface so you can observe and fine-tune your algorithm. You can adjust number_of_refresh_per_physical_second in the code to control simulation speed.
• cli.py runs the simulation without GUI for faster batch execution. It outputs output.csv in the data folder and logs in logs/.

Running the CLI simulation

Run:

python cli.py

The simulation may take a few minutes. At the end, it will report the total simulation time (e.g., 1,167,610 seconds) — this is the key metric you need to optimize and reduce. Note that this is the simulation time when system checks for all job completion condition. A precise time of the last job completion can be found in end_time column of output.csv file.

Submission Guidelines

When you are ready to submit your solution, please prepare the following:

1. YouTube Presentation video explaining your solution.
2. PowerPoint Slides supporting your presentation.
3. Source Code:
   • Modified job_planner.py plus any supporting scripts that depend only on job_planner.py.
   • Other modifications outside of this are not allowed.
   • Zip these files into Codes_TeamName.zip.
   • Include a README.txt describing how to run your code and integrate it into the existing codebase.
4. Simulation Results:
   • data/output.csv
   • the latest log file from logs/ demonstrating your improved results.
5. Team Photos.

---

If you have any questions or need assistance, feel free to reach out. Good luck and happy coding!