AIIDE 2019 Artifact

• Final paper: https://github.com/GAIGResearch/java-pommerman/raw/master/AIIDE-19_paper-46.pdf

• Download the artifact: clone this repository (master branch) or download as .zip file here: https://github.com/GAIGResearch/java-pommerman/archive/master.zip

• Instructions to run the artifact:

0. You need Java 8.0.1 version (or higher) to run this code.
1. Option 1: download the source, compile and execute the class Run.java (see below for execution modes).
2. Option 2: directly run run.jar, included in jars/

Executing Run.java / run.jar

This runs either a single game of pommerman (visuals on) or a series of games (headless), reporting statistics at the end. The usage is 'java Run' or 'java -jar run.jar' with 8 parameters:

• [arg index = 0] Game Mode. 0: FFA; 1: TEAM
  
• [arg index = 1] Number of level generation seeds [S]. "-1" to execute with the ones from the paper (20).
  
• [arg index = 2] Repetitions per seed [N]. "1" for one game only with visuals.
  
• [arg index = 3] Vision Range [VR]. (0, 1, 2 for PO; -1 for Full Observability)
  
• [arg index = 4-7] Agents. When in TEAM, agents are mates as indices 4-6, 5-7:
  
  • 0 DoNothing
    
  • 1 Random
    
  • 2 OSLA
    
  • 3 SimplePlayer
    
  • 4 RHEA 200 itereations, shift buffer On, pop size 1, random init, length: 12
    
  • 5 MCTS 200 iterations, length: 12
    
  • 6 Human Player (controls: cursor keys + space bar)
    

Examples:

• A single game with full observability, FFA. This is also the default mode when no arguments are passed:
  • java -jar run.jar 0 1 1 -1 2 3 4 5
• A single game with partial observability, FFA, where you're in control of one player:
  • java -jar run.jar 0 1 1 2 0 1 2 6
• Executes several games, headless, FFA. Two different random seeds for the level generation, repeated 5 times each (for a total of 5x2 games).
  • java -jar run.jar 0 2 5 4 2 3 4 1
• Executes several games, headless, TEAM, repeated 10 times each. Same configuration as the one used in the paper, including the 20 seeds.
  • java -jar run.jar 1 -1 10 4 5 3 5 3

Notes:

• If you provide N=1, the program will run a single game, with graphics on, using the agents specified in parameters 4-7.
• If you provide S=-1, the program will run N games with the specific 20 seeds used in the AIIDE 2019 paper (graphics off, results reported at the end).
• If you provide any other S>1, the program will run N games with S random seeds (total games, NxS), graphics off, using the agents specified in parameters 4-7 and results being reported at the end.
• The Human Player (option 6) is only available when N=1.

You can modify the code to execute different games as well (i.e. different agents or their parameters). For extra Java-pommerman wiki/documentation, visit this: https://github.com/GAIGResearch/java-pommerman/wiki

All games you play are logged in res/gamelogs/

Extra

All raw data and plots for the results reported in the paper is in the data/ folder (contains more plots than what could be included in the paper). You can either:

1. Download the Zip file here https://github.com/GAIGResearch/java-pommerman/blob/master/data/Analysis%20of%20Statistical%20Forward%20Planning%20methods%20in%20Pommerman.zip?raw=true
2. Explore the raw data and all generated plots on the repository yourself: https://github.com/GAIGResearch/java-pommerman/tree/master/data/Analysis%20of%20Statistical%20Forward%20Planning%20methods%20in%20Pommerman