CS7IS2_GroupProject

This is a GitHub repository which codes and analysis are submitted in partial fulfilment of the requirements for Trinity College Dublin Computer Science module CS7IS2 Artificially Intelligence.

Participant:

• Fan Bu, [email protected]
• Xin Lyu, [email protected]
• Xin Shu, [email protected]
• Yanxiang Chen, [email protected]

Environment and packages

The version listed below are used, developed and tested.
Using other versions may or may not compile the scripts.
Packages that are not listed in here do not require specific version.

• Python 3.6.13 or higher
• numpy 1.22.3
• pygame 2.1.2
• tensorflow:
  • On Intel and NVidia core: 1.15.5 (CPU-only or accelerated by CUDA v11.x & cuDNN v8.3.x)
  • On AMD Ryzen and Radeon core: tensorflow-directml 1.15.5, but require version of Python to be not higher than 3.8.*.

Addressed and solved game

Sudoku is a logic-based, combinatorial number-placement puzzle. In classic Sudoku, the objective is to fill a 9 × 9 grid with digits so that each column, each row, and each of the nine 3 × 3 subgrids that compose the grid (also called "boxes", "blocks", or "regions") contain all the digits from 1 to 9. The puzzle setter provides a partially completed grid, which for a well-posed puzzle has a single solution.

Usage

Clone git repository:

git clone https://github.com/Xin-Shu/CS7IS2_GroupProject.git

With an installed conda environmnet or python virtual environment:

pip install ./requirement.txt

Compile sudoku game:

python sudokuGame.py

To use algorithms on the game:

• Press q to set an Easy level sudoku puzzle with up to 35 blanks.
• Press w to set a medium level sudoku puzzle with up to 41 blanks.
• Press e to set a hard level sudoku puzzle with up to 52 blanks.
• Press a to crack the game using algorithm AC-3 solver.
• Press s to crack the game using algorithm Backtracking.
• Press l to crack the game using algorithm Deep Neural Network, with a given pretrained model.
• Press g to crack the game using algorithm Genetic Solver.

Used AI Algorithms

• AC-3 Backtracking, average solving time: 48 msec
• Genetic Algorithm, average solving time: 76 msec
• Deep Neural Network, average solving time: 1160 msec

Links

• Link to our GitHub repository:

github.com/Xin-Shu/CS7IS2_GroupProject

• Link to our Overleaf report:

overleaf.com/read/mpsjcjhrgcqp

• Links to our presentation video:

drive.google.com/file/d/13ugbmA0oZOj8Vhty6x3yIHxstCvHyJrI/view?usp=sharing

https://github.com/Xin-Shu/CS7IS2_GroupProject/blob/main/Presentation.zip

Contributions

• Bu Fan: AC-3 algorithm development, report writing.
• Xin Lyu:baseline (Depth-first Search) algorithm development, report writing.
• Xin Shu: proposal and development of CNN solution, refine the sudoku game scripts from forked repositories, settle project GitHub repository, voice over the presentation video, report writing.
• Yanxiang Chen: development of Genetic Algorithm, report writing, compatibility checking of codes in different envs.

Referred GitHub repositories or websites

• Sudoku game code:
  • puzzle generating: github.com/sarthak1905/sudoku-game-python
  • GUI display: geeksforgeeks.org/building-and-visualizing-sudoku-game-using-pygame/
• AC-3 algorithm: github.com/stressGC/Python-AC3-Backtracking-CSP-Sudoku-Solver
• CNN architecture: github.com/Kyubyong/sudoku
• Genetic algorithm:github.com/chinyan/Genetic-Algorithm-based-Sudoku-Solver