README.md

Labyrinth Intelligence & Analysis

Multi Agent System + Data Mining on Simulations

Project developed for Agents and Distributed Artificial Intelligence @FEUP, by @Ross65536, @msramalho and @jflcarvalho

Multitude of agents (and agent types) looking for a way out...

Instructions

To build:

$ ./gradlew build

To run:

# in project root
$ ./gradlew run

This will run with a stack size of 512MiB.

To run with a configuration file (YAML or JSON), specify the file path as the first argument:

$ gradle run --args="configurations/agents.yaml"
# or
$ gradle run --args="configurations/defaults.json"

To package

Simply run:

$ ./gradlew jar

This will create a <name>.jar file in the ./build/libs directory.

This is a dependency free .jar which means you only need the .jar and a java runtime to execute it since all dependencies and resources are already included in the zip.

You can then run

$ java -jar <path to jar>/<name>.jar [configuration file]

Or with a custom Stack size

$ java -Xss<MiBAmount>m -jar <path to jar>/<name>.jar [configuration file]

Configuration Example

Configuration File

YAML:

seed: 1
batchMode: true
mazeHeight: 50
mazeLength: 50
numForwardAgents: 1
numBacktrackAgents: 0
numRandomAgents: 0
numNegotiatingAgents: 5

CLI arguments

You can add "any" (if synchronized) arguments to the CLI that you would to the configuration file. To add a seed you can add an argument -seed=13 for a seed of 13. If you supply only one argument without a flag - a config file is used. If you supply a flag the arguments are used.

Full Example:

$ gradle run --args="-batchMode=true -mazeHeight=30 -mazeLength=30 -seed=1 -numForwardAgents=2 -numBacktrackAgents=2 -numRandomAgents=2 -numNegotiatingAgents=2 -numSwarmAgents=0 -slownessRate=10 -statisticsPath="metrics/new.csv""

Scripts

To run the script with the experiments run for example:

$ python scripts/big-experiments.py build/libs/aiad-labyrinth-0.1-SNAPSHOT.jar y