Incorporating Fairness into Agent Interactions Modeled as Two-Player Normal-Form Games

Abstract

Many multi-agent interaction problems, like auctions and negotiations, can be modeled as games. Game theory, a formal tool for game analysis, thus can be used to analyse the strategic interactions among agents and facilitate us to design intelligent agents. Typically the agents are assumed individually rational consistent with the principle of classical game theory. However, lots of evidences suggest that fairness emotions play an important role in people’s decision-making process. To align with human behaviors, we need to take the effects of fairness motivation into account when analysing agents’ strategic interactions. In this paper, we propose a fairness model which incorporates two important aspects of fairness motivations, and the solution concept of fairness equilibrium is defined. We show that the predictions of our model successfully reflect the intuitions from both aspects of fairness motivations. Besides, some general results for identifying which outcomes are likely to be fairness equilibria are presented.