Non-Bayesian social learning model with periodically switching structures.

Area of Science:

• Social learning dynamics
• Network theory
• Probability theory

Background:

• Non-Bayesian social learning models often require strongly connected networks for convergence.
• Existing models lack flexibility in handling dynamic network structures.

Purpose of the Study:

• To investigate the convergence properties of a non-Bayesian social learning model with periodically switching structures.
• To relax the stringent connectivity requirements typically imposed on such models.
• To provide convergence rate estimations for successful social learning.

Main Methods:

• Mathematical and rigorous validation of model dynamics.
• Analysis of convergence under relaxed network configurations.
• Numerical demonstrations using representative examples with switching structures.

Main Results:

• The model converges to a true state of social learning even with non-strongly connected networks during switching periods.
• Convergence rate estimations for successful social learning were established.
• Analytical conditions and estimations were numerically validated.

Conclusions:

• The proposed model offers a more flexible framework for studying social learning dynamics.
• Results suggest applicability to understanding complex real-world social activities.
• Relaxed network conditions do not impede convergence in this non-Bayesian social learning model.