The Virtual World of Decision Making Turtles is implemented in Director 7, a product of Macromedia. The demos are Shockwave movies and require a browser that supports this file type. Download the Shockwave plug-in if needed.

Demos have a file size of ~ 270k

Basic Demo: Implementation as laid out on the Description Page.

Goal Based Demo: Along with the basic implementation, this demo include 9 red patches. These red patches are considered "goals" for the DMTs to reach. The purpose of these red goals is to attempt to evolve the population based on a criterion not directly related to a DMTs survival. Along with the environment becoming more complex, a DMTs genome must become more complex and lengthy. This slows the evolution of the population while introducing more diversity. An additional statistic for "The Average X's Collected" (X's marks the spot) represents the average number of red patches a DMT has reached. This number has a maximum possible value of 9, because 9 red patches appear in the environment. It has not yet been determined if these DMTs can evolve to become good "X finders". The genetic algorithm, the period between generations and the length the simulation is run for may have to be tweaked to achieve the desired results.

Evolution of Rebounding Demo: You may notice in the first two demos that some of the DMTs make it outside of the confines of the patches. This is due to a lack in accuracy in my rebounding function. This function's purpose is to turn the DMTs around when they reach an edge, but some of the DMTs just keep turning back, determined to leave the patches. I decided I should fix this and so set out to improve the rebounding function when I thought of an even better solution; make the DMTs evolve rebounding. This simulation attempts to demonstrate this evolution of advanced behaviors from simple actions. A few modification are made to the basic demo to help encourage a rebounding behavior: (1) A ring of "barren" patches line outer edge of the board, these patches can grow no food. This gives the DMTs time to decide to turn towards greener pastures before leaving the world. (2) The genetic algorithm considers DMTs that die off the board less desirable than those that die on the board. (3) The statistic "Number died off board" is added to the statistics area. The first few generation may have a large number of DMTs who run right off the board, but this number should decrease as the population evolves.

For versions of this project that can be run on your own computer, along with documentation, see the Downloads page. These versions include the ability for the simulation to save data to files on your hard drive, allowing the simulation to pickup where it left-off when it was last halted.