EPFL CS 430 Lab and project programmed in java
Build a pick and delivery simulation system on top of RePast.
The space is a 2D Torus containing randomly genrated food.
The agent has Movement, Reproduction, Eating and Dying action.
Run the program by
java -jar ./hw0/out/artifacts/lastname1_lastname2_in_jar/lastname1-lastname2-in.jar "" false
The initial status of the world including gridSize, numInitRabbits, numInitGrass, grassGrowthRate and birthThreshold can be set in the GUI generated.
This lab builds a reactive agent using value iteration to find an optimal strategy for a pickup-delivery problem, which is a instance of Markov decision process.
The detailed descriptions of the problem can be found at here.
To try different settings, please follow the instructions here to create or modify the xml files.
For visualing how our agent performs, run
java -jar ../logist/logist.jar config/reactive.xml reactive-random
This lab builds a centralized agent.
The project can be run with
java -jar ../logist/logist.jar config/centralized.xml centralized-random
This lab builds a deliberative agent.
The project can be run with
java -jar ../logist/logist.jar config/deliberative.xml deliberative-random
This lab is builds an auction agent.
Our implementation got the third place in the final contest of 2019 Fall.
To run the tournament:
Put the jar files of your agents in a folder, e.g. ./agents
Create a tournament:
java -jar './logist/logist.jar' -new 'tour' './agents'
Run the tournament:
java -jar './logist/logist.jar' -run 'tour' './config/auction.xml'
Save the results:
java -jar './logist/logist.jar' -score 'tour'
inside the reative directory.