Cooperation in collective dilemmas under opinion-based risk perceptions

Collective risk dilemmas are evolutionary games in which every player can contribute some amount to avoid a certain risk of failure. The main goal of this study is to integrate, within a collective dilemma, the influence, evolution, and formation of the perceived risk by individuals of the population. In order to understand the effects of subjective evolving opinions about risk perception in the evolutionary game, we pair a traditional collective game model of homogeneous groups with a network of players evolving their opinions or perceptions about the risk of common failure. We study the evolution of the players’ perception about the risk and how different network topologies and opinion models, with and without considering the outcome of the game, affect the output of the evolutionary game. We show that cooperation generally increases when the evolution of opinions leads to consensus under unimodal and polarized initial opinions, for all the evaluated scenarios. Even when the population has similar mean final opinions, the transition of the opinions affects the final cooperation of the game. These findings highlight the practical relevance of peer opinion exchange between agents in real dilemmas as a mechanism to increase cooperation and avoid collective failures.