https://github.com/heziode/tron-multiplayer
A Tron version for AI experiments (no GUI)
https://github.com/heziode/tron-multiplayer
Last synced: about 1 year ago
JSON representation
A Tron version for AI experiments (no GUI)
- Host: GitHub
- URL: https://github.com/heziode/tron-multiplayer
- Owner: Heziode
- Created: 2017-06-10T11:49:48.000Z (about 9 years ago)
- Default Branch: master
- Last Pushed: 2017-06-17T14:49:45.000Z (almost 9 years ago)
- Last Synced: 2025-01-30T18:38:42.223Z (over 1 year ago)
- Language: Java
- Size: 55.7 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Coalitions of artificial intelligence for the game Tron
This project was designed for a university project.
The objective of this project it's to determine if exist in configuration who a player can win against a coalition of other players.
> This project is basically designed for a "statistical study". It is built to be run easily from a script. **There is no GUI**
>
> **This project is designed to run on Linux and require JavaFX.**
For this project, I implemented the Paranoid algorithm, a multiplayer variant (more than two players) of MinMax. It assumes that all players form a coalition against a single player:
```
Algorithm paranoid (ei,d)
if d = 0 then
return f(ei)
else
if ei is a player node then
m ← −∞
for all ej ∈ successors(ei) do
m ← max(m, paranoid(ej,d−1))
end for
else
m ← +∞
for all ej ∈ successors(ei) do
m ← min(m, paranoid(ej,d−1))
end for
end if
return m
end if
```
## Build project
```
git clone https://github.com/Heziode/tron-multiplayer.git
cd tron-multiplayer
ant
```
## VM parameters
Mandatory :
- tron.coal : integer value (separated by spaces): depth of calculation of each player coalitions
- tron.nbCoal : integer value (separated by spaces): number of players in each coalition
Optional :
- tron.gridSize : integer value: grid size (square grid), by default, randomly depending on the number of players
- tron.gui : boolean, true or false: indicates whether the program should be started with or without GUI (True by default)
- tron.ptrace : boolean (false by default): true to save the paranoid calculation tree in DOT, false otherwise
- tron.player.order : integer value (separated by spaces): order of the players by coalitions (eg 0 1 0 2 = place one first of a coalition player 0, then a coalition player 1, etc.)
- tron.player.position : integer value (0 default): position (in the player list) of the first player to play.
- tron.player.shift : integer value (0 by default): shifts the location of the first player to the grid border
- tron.player.location : integer pair (separated by spaces): coordinates "x,y" of each player
- tron.result : boolean (false by default): true to save the result of the game to a file (in CSV format), false otherwise
- tron.result.name : string: name of the result file. The default is a random UUID (Universally Unique IDentifier).
- tron.random : boolean (false by default): true to enable player placement, grid size and first player to play randomly
- tron.result.print : boolean (false by default): true to display the result of the part (at the CSV foramt) in the terminal, false otherwise
Exemple : **java -jar -Dtron.gui=false -Dtron.coal="5 3" -Dtron.nbCoal="1 2" -Dtron.gridSize=7 Tron_Multiplayer.jar**
This will launch a game without a graphical use interface (GUI) and a solo player with a depth of calculation of 5 against a coalition of 2 players with a depth of calculation of 3 on a grid of 7 by 7.
## Export results of a game in CSV format
It is possible to export the result of a game in a CSV format using the parameter:
`-Dtron.result=true`
The only separator character used is the comma: **,**
> Results are saved only in non-GUI mode
The export format obtained is as follows (with an example of result in the second line):
| Size of the grid | Winning coalition | Number of strokes | First player index | Total number of players | Number of coalitions | Location of players at the start | Number of players per coalition | Research depth of coalitions |
| :--------------: | :---------------: | :---------------: | :----------------: | :---------------------: | :------------------: | :------------------------------: | :-----------------------------: | :--------------------------: |
| 7 | 1 | 37 | 0 | 3 | 2 | \[0;0\] \[6;2\]\[2;6\] | 1 2 | 5 3 |
## Export Paranoid calculation
It is possible to export in DOT language the calculations made by paranoid to each stroke using the parameter:
`-Dtron.ptrace=true`
> It is possible to generate SVG to visualize the DOT language. This requires, on the other hand, to have the utility **dot**
You can use the **toDot.sh** script to generate them. It assumes that the files are located in a subdirectory **paranoid_traceback**.
Legend:
- **V**: Square visited by a player
- **J**: Square occupied by a player
- **T**: Square belonging to the territory of a player
- **B**: Square blocked, which is and will never be accessible during the game
- **R**: Value of this state of games according to the player who plays at *node0*
- **node**: the node numbers are given according to the paranoid calculation order
- **Red arrow**: best solution found by paranoid
- **MAX**|**MIN**: indicates whether the node takes the maximum or minimum value of its wires
## Massive execution
You can make mass executions of the program through the script contained in the script directory.
> Scripts are written in JavaScript and require Node.js
## License
MIT