Using Evolutionary Computation to Find Parameters that Promote Egalitarian Major Evolutionary Transitions

Evolutionary transitions, where replicating units combine to form more complex units, are a major source of the complexity found in nature. In this paper we aim to find conditions that promote egalitarian major transitions in a digital artificial ecology. We identify major transitions in this context by observing changes in fitness across different levels of organization. Fitness increases primarily at the community level suggest the occurrence of major transitions. We employ a genetic algorithm using lexicase selection to find regions of parameter space that promote community-level fitness increases. This approach successfully finds multiple ecological community structures that appear to support major transitions. These results illustrate the power of evolutionary computation for exploring the parameter space of complex simulations and push us closer to an understanding of the factors that lead to egalitarian major transitions.