READ ME:

ChainReactionAI

A chain reaction game based on AI

References:

1. Minimax Algorithm - Depth Limited : 2 since the state space for th game is large
2. Numpy Documentation
3. Matplotlib Documentation
4. TextWrap Documentation

Dependencies Used:

For Chain Reaction the following are utilized

1. Python - version 3.8
2. Numpy
3. Pytorch
4. Matplotlib
5. TextWrap

Links

References used from :

Author: Katz, Garrett Topic : Minimax Game Trees, Neural Networks Syracuse University https://colab.research.google.com/drive/1YR8HjSw8K0n684S_oGnZPpU69SmOw355?usp=sharing#scrollTo=N6kj91QAfbuW https://colab.research.google.com/drive/1JhOppwXwm47yk-AK7y7L5WTaaNDgCWXD?usp=sharing#scrollTo=TtHr_xOA7fa7

How to run the program?

STEP 1: Clone / Download code from our Github repo and decompress the file. STEP 2: Open terminal on user’s system. STEP 3: Go to the ChainReaction folder and run the “chain_reaction.py” ($ python chain_reaction.py). STEP 4: Select if you want to play the game with a Human, Baseline AI or Tree Based AI by inputting the corresponding number. Implementation of Tree Based AI is in progress STEP 5: Play the game!

Since the game is interactive based on graphical displays, a Windows based machine would be ideal to run the game. This has been tested using the following configurations: OS: Windows IDE: Python IDLE 3.8 64x RAM: 16 GB Disk Space: 250KB

IMPORTANT NOTE!

If the game is being run on a Windows system, make no changes to the code. If the game is being run on a Macintosh System, make the following changes in ‘print_board.py’ under display. Comment the lines plt.ion(), plt.pause(5) and plt.close() except plt.show(block = False).

TreeAI +NN = We have tried several configurations of layers for input to_hidden and hidden to_output, however we are currently unable to completely process the request. We are able to generate the training and testing examples with augmentation which increases the size from 100 to 800 scenarious. The depth for minimax is considered 2 since the state space is very large and we have constraint on recursive limit and higher depts might lead to exceeding the limit. Partial output is availale due to kernel size error. Please do not use this option for game or simulation untill updates are available.

Enjoy the Game !