First published 2 September 2013
Evolving gene regulatory networks controlling foraging strategies of prey and predators in an artificial ecosystem
Joachim Erdei, Michal Joachimczak and Borys Wróbel
Co-evolution of predators and prey is an example of an evolutionary arms race, leading in nature to selective pressures in positive feedback. We introduce here an artificial life ecosystem in which such positive feedback can emerge. This ecosystem consists of a 2-dimensional liquid environment and animats controlled by evolving artificial gene regulatory networks encoded in linear genomes. The genes in the genome encode chemical products which regulate other genes, sense the environment (the scent of food, prey and predators), control the animat's movement, and its foraging strategy. An animat can switch multiple times in its life between two foraging strategies (with different metabolic costs): a predator can derive food from the prey, prey just from food that diffuses in the environment. When an animat consumes enough food (or prey), it produces an offspring with a mutated genome. Mutations introduce variation into the population, and this diversity together with selective pressures leads to the evolution of control for diverse foraging strategies in an ecosystem that can support hundreds of individuals.