Stability of mixed-strategy-based iterative logit quantal response dynamics in game theory

Area of Science:

• Game Theory
• Behavioral Economics
• Computational Economics

Background:

• Static quantal response equilibrium (QRE) is a foundational concept in game theory.
• Existing models often lack dynamic stability analysis for equilibrium predictions.

Purpose of the Study:

• To extend the static quantal response equilibrium (QRE) into a dynamic, iterative evolution process.
• To introduce and define stable quantal response equilibriums (SQREs) and unstable quantal response equilibriums (USQREs).
• To compare the predictive power of SQREs against Nash equilibriums (NEs) and QREs using experimental data.

Main Methods:

• Utilized the Logit quantal response form as the response function within an iterative dynamic process.
• Classified fixed points of the dynamic process into stable (SQREs) and unstable (USQREs) equilibriums.
• Analyzed the relationship between SQREs, NEs, and QREs.

Main Results:

• QREs emerge as fixed points within the dynamic process.
• SQREs and USQREs are identified based on their long-term stability.
• Preliminary comparisons indicate that SQREs exhibit superior predictive accuracy compared to QREs and NEs for specific game types.

Conclusions:

• The dynamic extension provides a more nuanced understanding of equilibrium concepts in games.
• SQREs offer a potentially more robust solution concept for predicting behavior in strategic interactions.
• Further research is warranted to explore the applicability of SQREs across a wider range of games and experimental settings.