Cooperation, competition and the emergence of criticality in communities of adaptive systems

TL;DR

This paper models how communities of adaptive agents tend to self-organize near critical points, balancing order and disorder, with cooperation and competition influencing this criticality.

Contribution

It introduces a model based on information theory and statistical mechanics showing how agent communities converge to critical states and analyzes the effects of cooperation and competition.

Findings

Criticality emerges as the optimal state for adaptive communities.

Competition promotes convergence to criticality.

Cooperation tends to hinder criticality, leading to more ordered or disordered states.

Abstract

The hypothesis that living systems can benefit from operating at the vicinity of critical points has gained momentum in recent years. Criticality may confer an optimal balance between exceedingly ordered and too noisy states. We here present a model, based on information theory and statistical mechanics, illustrating how and why a community of agents aimed at understanding and communicating with each other converges to a globally coherent state in which all individuals are close to an internal critical state, i.e. at the borderline between order and disorder. We study --both analytically and computationally-- the circumstances under which criticality is the best possible outcome of the dynamical process, confirming the convergence to critical points under very generic conditions. Finally, we analyze the effect of cooperation (agents try to enhance not only their fitness, but also that of other individuals) and competition (agents try to improve their own fitness and to diminish those of competitors) within our setting. The conclusion is that, while competition fosters criticality, cooperation hinders it and can lead to more ordered or more disordered consensual solutions.