Welfare and price of anarchy under incomplete information in network games

Develop a welfare analysis and characterize the price of anarchy for the quadratic network game u_i(x) = - (1/2) γ_i x_i^2 + (β_i + ∑_{j ≠ i} g_{ij} x_j) x_i when agents have incomplete information about the parameters γ_i, b_i, and g_{ij} and reason within standard Bayesian frameworks, quantifying inefficiency and deriving tractable characterizations in this information environment.

Background

The price-of-anarchy characterization in the paper assumes complete information about utility parameters. The author argues this is unrealistic and observes that the linear structure should permit tractable Bayesian analysis.

Despite this tractability, the author states that the paper of welfare and price of anarchy under incomplete information is completely open, pointing to a gap in the literature at the intersection of network game theory and Bayesian uncertainty.