Report 4 - HexKitties/CMPUT355HexKitties GitHub Wiki
Fourth meeting:
Meeting Schedule:
Date: Nov 17 2020
Start Time: 10:00 pm
End Time: 11:00 pm
Attendants: Everyone
What We Discussed:
- Wait for how long when generating the next move? (10 sec)
- Main trained board size: 5x5 (different board size can be trained, but we will stick with this for now)
- Bugs:
- does not update display properly when AI player is figuring out the next step
- stuck in the program when Monte Carlo is searching for the next step (probably print a loading message for better user experience)
- test algo's performance (currently is using 10 seconds for Monte Carlo search, and the choices made are satisfying so far)
- starting player changed for the next game (default starting player should always be red)
- player mode is changed back to real player mode when a game ends
- undo button bug in AI player mode (should undo the AI's last move too)
Work distribution(merging with hex_v3.py):
Member | Tasks |
---|---|
Winning message, Draw winning path (blink) | Youwei |
Optimize current Monte Carlo simulation, Performance: compete with other players (winning probability) | Siyuan & Jiaxin |
Store new trained data in a file, Retrieve trained data | Yuxin |
fix mode changing bug, fix update display bug, fix next starting player bug | Jeremy |
Later
- “Dead” bug: stuck in the program for 10 secs (loading...)
- user gets to choose Monte Carlo thinking time
Next meeting time:
Sunday, Nov 22: 10:00 pm