Dynamic landscape models of coevolutionary games

TL;DR

This paper introduces dynamic landscape models to analyze coevolutionary games, allowing for the examination of strategy and network updates using landscape measures, and relates these to game dynamics like fixation probabilities.

Contribution

It presents a novel application of dynamic fitness landscapes to coevolutionary games, linking landscape properties with game dynamics and network features.

Findings

Landscape measures correlate with fixation probabilities.

Network properties influence landscape ruggedness.

Dynamic landscapes provide new insights into coevolutionary processes.

Abstract

Players of coevolutionary games may update not only their strategies but also their networks of interaction. Based on interpreting the payoff of players as fitness, dynamic landscape models are proposed. The modeling procedure is carried out for Prisoner's Dilemma (PD) and Snowdrift (SD) games that both use either birth--death (BD) or death--birth (DB) strategy updating. The main focus is on using dynamic fitness landscapes as a mathematical model of coevolutionary game dynamics. Hence, an alternative tool for analyzing coevolutionary games becomes available, and landscape measures such as modality, ruggedness and information content can be computed and analyzed. In addition, fixation properties of the games and quantifiers characterizing the interaction networks are calculated numerically. Relations are established between landscape properties expressed by landscape measures and quantifiers of coevolutionary game dynamics such as fixation probabilities, fixation times and network properties.