Games with Rational and Herding Players

TL;DR

This paper introduces a new equilibrium concept for large population games with mixed rational and herding players, revealing how herding behavior can unexpectedly improve system efficiency and social welfare.

Contribution

It develops the alpha-Rational Nash Equilibrium framework and analyzes the impact of herding on game outcomes, including efficiency and social welfare, in various network scenarios.

Findings

Herding can lead to the emergence of new equilibria.

Herding behavior can improve overall system efficiency.

Adding links can increase efficiency when herding is present.

Abstract

Classical game theory is a powerful framework to analyze the strategic interactions among rational players. However, in many real-life scenarios, players choose actions based on their inherent natural tendencies rather than deliberate reasoning. In this paper, we develop an analytical framework to study large population games with an alpha-fraction of rational and (1-alpha)-fraction of herding players. We introduce a new notion of equilibrium called alpha-Rational Nash Equilibrium (in short, alpha-RNE) and discuss its interpretations. Some classical equilibria may disappear, and some new ones may emerge, but only for smaller alpha >0. Interestingly, rational players benefit from the presence of herding and may even achieve utility exceeding the socially optimum. Even more strikingly, in some cases, the herding players also benefit, attaining utility close to the social optimum. We further study the effect of the herding fraction on system performance using measures such as the Price of Anarchy (PoA). In transportation networks, a well-known paradox first studied by Pigou and later by Braess typically arises from rational decision-making: adding an extra link can reduce overall efficiency. Our analysis leads to a different conclusion. When a substantial fraction of users exhibit herding behavior, introducing a new link can increase efficiency, provided herding choices can be suitably influenced. The gains are larger when the herding fraction is higher and/or congestion is lower. By contrast, when herding decisions cannot be influenced, the added link may become detrimental. We also study a bandwidth sharing game in which herding tendencies improve system efficiency. Finally, we discuss the mechanism or influence design in the presence of herding, highlighting both opportunities and risks.