Objective/Goal

This project presents a simulator for modeling an environment containing intelligent herbivorous organisms, and their evolution inside it. In addition to this, the simulator is capable of changes in plant availability, which can serve as a model for climate change.

Methods/Materials:

Each of the simulated organisms' brains are modeled by a feedforward neural network, which evolves as part of the organisms genome. This occurs as they mate and reproduce, recombining with a chance of mutation. The organisms are capable of viewing their environment and fellow members of the population, to allow them to decide how to interact. Each organism consists of sensors (eyes), brain, and body.

The simulation, written in JavaScript and HTML5, consists of a habitat, represented as a two dimensional circlular area, containing organisms and plants.

Results

I have successfully built a configurable evolutionary simulator, and used it to quantify the effects of climate change on simulated herbivore populations.