Team Members: 孫宜君、楊竺耘、陳宥安
Interface reference: Minesweeper

• Algorithm we used
  • Straightforward Algorithm
    • 詳見我們製作的Notion筆記
  • Expected value of being mines
    • 首先遍尋周圍的點，並且將這次遍尋的點的出現頻率紀錄在cell-freq中，接著再將這些點的周圍也遍尋一次，計算出周圍沒有走過的點的數量，紀錄於neighbor-num;將該點的周圍地雷數除以neighbor-num，並且採持續累加，最後再除以該點出現的頻率，如此便計算出該點的期望值。
• Strategy
  • Straigntforward Algorithm + Expected value
    • Judge the safe spots by Straightforward Algoreihm
    • Calculate the expected value of the grids where are connected to the opened spots.
    • If there is safe spot, open it (randomly choose one if there are multiple safe spots).
    • Otherwise, choose the spot with lowest expected value (randomly choose one if there are multiple spots that have same expected value).

• Package you need:
  • pygame
• You can download the needed package by

  pip3 install -r requirements.txt
  

  or

  pip3 install pygame
  

• You can start the game by running:

  python3 runner.py
  

  and press the button "Play Game". Then you will see the game interface.

• Manual Mode
  • Simply start the game by clicking the grid with left mouse button.
  • You can mark the grid by clicking the grid with RIGHT mouse button. Pressing RIGHT mouse button again can numark the grid.
• AI move
  • The AI we wrote will move a step for you based on the knowledge database the AI abtained.
• Autoplay
  • The AI we wrote will play a round for you and display the result (win or loss).
• Test
  • The AI we wrote will run 10000 rounds for you and display the following infomation:
    • Won rate
    • Average step (if loss)
    • Total random step
    • Average score
    • Average score (if loss) you can change the time of running round in runner.py.

• Video link: https://www.youtube.com/watch?v=Gzj6q7J6lxc&ab_channel=%E9%99%B3%E5%AE%A5%E5%AE%89