Population dynamics and games of variable size

TL;DR

This paper introduces Variable Size Game Theory (VSGT), where players decide game size strategically, and explores its implications for cooperation, speciation, and epidemic modeling, demonstrating its potential to explain complex biological and social phenomena.

Contribution

The paper develops VSGT, a novel framework allowing players to choose game size, and applies it to diverse scenarios including cooperation, speciation, and epidemic dynamics.

Findings

Large groups can dominate costly tasks, supporting eusociality evolution.

Different strategies lead to spontaneous population segregation, modeling sympatric speciation.

VSGT can recast epidemic models like SIRS into a game-theoretic framework.

Abstract

This work introduces the concept of Variable Size Game Theory (VSGT), in which the number of players in a game is a strategic decision made by the players themselves. We start by discussing the main examples in game theory: dominance, coexistence, and coordination. We show that the same set of pay-offs can result in coordination-like or coexistence-like games depending on the strategic decision of each player type. We also solve an inverse problem to find a $d$-player game that reproduces the same fixation pattern of the VSGT. In the sequel, we consider a game involving prosocial and antisocial players, i.e., individuals who tend to play with large groups and small groups, respectively. In this game, a certain task should be performed, that will benefit one of the participants at the expense of the other players. We show that individuals able to gather large groups to perform the task may prevail, even if this task is costly, providing a possible scenario for the evolution of eusociality. The next example shows that different strategies regarding game size may lead to spontaneous separation of different types, a possible scenario for speciation without physical separation (sympatric speciation). In the last example, we generalize to three types of populations from the previous analysis and study compartmental epidemic models: in particular, we recast the SIRS model into the VSGT framework: Susceptibles play 2-player games, while Infectious and Removed play a 1-player game. The SIRS epidemic model is then obtained as the replicator equation of the VSGT. We finish with possible applications of VSGT to be addressed in the future.