BioANT - biologically plausible computer simulation of an environment with ants

Abstract

The system BioAnt, presented here, is an artificial life system that simulates an environment with ants. Each ant moves inside the environment on the basis of a biologically plausible artificial neural network, which employs the algorithm GeneRec for supervised learning. In order to obtain an initial configuration, a symbolic algorithm (based on a set of production rules) was created. The environment itself is three-dimensional and consists of the anthill, sugar, water, earth elevations, walls, and predators. The ants interact with the environment using the senses of smell, vision, touch, and hearing, always following the commands of the neural controller. The system aims to contribute to research that deals with development of artificial beings, i.e., intelligent agents, artificial life, robotics, and biologically plausible artificial neural networks. BioAnt is an implementation of typical behavioral aspects of ants in a computational simulation.