Discreet World

2007-04-15, download

This small-scale simulation program was inspired by the first few chapters of Artificial Intelligence: A Modern Approach. More specifically the cleaner agent of the vacuum world. My university studies did not include to much about agents that act on thier own. So I got curious enough to implement a synthetic world and develop an agent that tries to make it dust-free. I developed a map visualizer sub-system so you can enjoy the result in real time :) If you are interested in the source code you can find that in the downloadable jar file. The .jar file can also be run as a standalone application.


  1. First you have to select a map. The agent selector boxes appear. The map defines how many agents can you place on them, so it may vary
  2. Select a number of agents
  3. Hit the start button
  4. You can pause/resume the simulation any time. You can also add or remove agents after stopping the simulation.

Agents types

  • Vercy’s cleaner agent: This agent systematically traverses the map looking for dust. After the whole maps has been revised it returns to its start position
  • Dummy agent: chooses a random action regardless of the sensed data. It was a great help for debugging, and it looks kind of funny