Adversarial Behavior in Network Mechanism Design

This paper studies the effects of and countermeasures against adversarial behavior in network resource allocation mechanisms such as pricing and auctions. It models the heterogeneous behavior of users, which ranges from altruistic to selfish and even to malicious, using game theory. The paper adopts a mechanism design approach to quantify the effect of adversarial behavior and modify the mechanisms to respond. First, the \textit{Price of Malice} of the existing network mechanisms to adversarial behavior, which ranges from extreme selfishness to destructive maliciousness, is analyzed. Then, two methods are discussed to counter such adversarial behavior: one is a differentiated pricing to punish the malicious users and another is a detection method based on the expected utility functions of the ``regular'' users on the network. Finally, the results obtained are illustrated with multiple examples and numerical simulations.