Informational incentives for congestion games

We investigate the problems of designing public and private information disclosure mechanisms by a principal in a transportation network so as to improve the overall congestion. We show that perfect disclosure of information about the routes' conditions is not optimal. The principal can improve the congestion (i.e. social welfare) by providing coordinated routing recommendation to drivers based on the routes' conditions. When the uncertainty about the routes' conditions is high relative to the ex-ante difference in the routes' conditions (the value of information is high), we show that the socially efficient routing outcome is achievable using a private information disclosure mechanism. Furthermore, we study the problem of optimal dynamic private information disclosure mechanism design in a dynamic two-time step setting. We consider different pieces of information that drivers may observe and learn from at t = 1 and investigate qualitative properties of an optimal dynamic information disclosure mechanisms using numerical simulations.