The Impact of Bounded Rationality on Strategy Selection in the Prisoner’s Dilemma

Abstract

This study investigates the profound and often overlooked impact of bounded rationality on strategic decision-making, using the prisoner’s dilemma as a foundational model. A questionnaire issued by the author reports that over 70% of people prefer to cooperate rather than betray, which is completely different from what is predicted in theory. This stark empirical discrepancy not only challenges the core assumption of perfect rationality in classical game theory but also underscores the critical need to incorporate behavioral elements for accurate prediction of real-world outcomes. Bounded rationality is an important part that is always ignored by game theorists when analyzing cases. Although an increasing number of papers are describing the causes of bounded rationality, there is still a gap about how to optimize bounded rationality and prevent possible loss from being irrational. To bridge this gap, this study first analyzes the core causes and dualistic effects (both positive and negative) of bounded rationality, supported by real-world case studies. It then proposes and evaluates a series of practical optimization strategies—including a hybrid stochastic approach and a minimax regret framework—designed to leverage the cooperative potential of bounded rationality while insulating decision-makers from its associated risks.