Towards Eusociality Using an Inverse Agent Based Model

TL;DR

This paper presents an inverse agent-based model to discover behaviors leading to eusociality, highlighting how evolved colony efficiency relates to agent computational capacity and phenotypic plasticity.

Contribution

It introduces a novel inverse modeling approach to identify agent rules that produce eusocial behaviors and explores the role of phenotypic plasticity and computational capacity in colony success.

Findings

Evolved colonies show phenotypic plasticity in response to environment.

Colony efficiency correlates with agents' computational capacity.

Emergent behaviors include eusocial and novel population dynamics.

Abstract

The emergence of eusocial species is both very rare in evolutionary history and results in remarkably successful species. By inverting an agent based model, agent rules are discovered that display behaviors characteristic of eusocial species as well as other behaviors that lead to unexpected population dynamics. By holding the agents' genome constant across the colony and allowing the agents' rules to evolve, the individual behaviors exhibit phenotypic plasticity in response to environmental cues. The phenotypically driven reduction of intrinsic growth rates and the emergence of non-reproducing phenotypes both demonstrate selection pressure at the colony (system) level. The efficiency of an evolved colony is shown to have a strong relationship to the computational capacity of the agents. Various other emergent behaviors, both eusocial and otherwise novel, are identified and discussed. A path forward to more capable eusocial populations and inter-colony evolution is outlined.