Avoiding Unintended Consequences: How Incentives Aid Information Provisioning in Bayesian Congestion Games

TL;DR

This paper investigates how signaling and incentives can be used in Bayesian congestion games to improve system performance, revealing that combined strategies can prevent information disclosure from worsening outcomes.

Contribution

It introduces bounds on the benefits of signaling in Bayesian congestion games and demonstrates how incentives can complement signaling to ensure non-detrimental information sharing.

Findings

Signaling can sometimes worsen system performance.

Combining signaling with incentives guarantees non-worsening of outcomes.

Numerical examples illustrate the impact of information and incentives.

Abstract

When users lack specific knowledge of various system parameters, their uncertainty may lead them to make undesirable deviations in their decision making. To alleviate this, an informed system operator may elect to signal information to uninformed users with the hope of persuading them to take more preferable actions. In this work, we study public and truthful signalling mechanisms in the context of Bayesian congestion games on parallel networks. We provide bounds on the possible benefit a signalling policy can provide with and without the concurrent use of monetary incentives. We find that though revealing information can reduce system cost in some settings, it can also be detrimental and cause worse performance than not signalling at all. However, by utilizing both signalling and incentive mechanisms, the system operator can guarantee that revealing information does not worsen performance while offering similar opportunities for improvement. These findings emerge from the closed form bounds we derive on the benefit a signalling policy can provide. We provide a numerical example which illustrates the phenomenon that revealing more information can degrade performance when incentives are not used and improves performance when incentives are used.