Design of Incentive Compatible Protocols for Wireless Networks: A Game Theoretic Approach
Proceedings IEEE INFOCOM 2006. 25TH IEEE International Conference on Computer Communications
Thesis Abstract: In this thesis work, we design rigorous and efficient protocols/mechanisms for different types of wireless networks using a mechanism design  and game theoretic approach  . Our work can broadly be viewed in two parts. In the first part, we concentrate on ad hoc wireless networks  and  . In particular, we consider broadcast in these networks where each node is owned by independent and selfish users. Being selfish, these nodes do not forward the broadcast packets. All
... existing protocols for broadcast assume that nodes forward the transit packets. So, there is need for developing new broadcast protocols to overcome node selfishness. In our paper , we develop a strategy proof pricing mechanism which we call Immediate Predecessor Node Pricing Mechanism (IPNPM) and an efficient new broadcast protocol based on IPNPM. We show the efficacy of our proposed broadcast protocol using simulation results. Protocols for wireless ad hoc networks suffer in network performance that includes large routing overhead, low throughput, and large end-to-end delay. In such networks, the issues of quality of service (QoS) are even more complicated because of the lack of reliable methods to distribute information in the entire network. The integration of heterogeneous wireless technologies can improve the network performance, thereby meeting the demands for different quality of service (QoS). Motivated by this interesting research problem, we concentrate on heterogeneous wireless networks in the second part of our thesis work. In particular, we try to answer some of the resource allocation problems that arise in these networks. We consider the situation where a mobile/wireless user has to perform some parallel applications where each parallel application can be split over different wireless networks for execution. The mobile user is having access to heterogeneous wireless networks provided by a set of non-cooperative and selfish service providers. These service providers charge for allocating the radio resource to the mobile/wireless users. Ultimately, the problem of the mobile user is to procure the radio resource from these service providers to perform its applications while minimizing the total amount for it to pay to the set of network service providers. In our paper , we design an optimal procurement mechanism to solve the problem of mobile user using game theory and mechanism design. Our solution satisfies important game theoretic properties such as Individual Rationality and Bayesian Incentive Compatibility. We now briefly elaborate on each part of our work to provide more details. Part-1: Truthful Broadcast Problem for Wireless Ad hoc Networks In many applications, in particular civilian applications, of wireless ad hoc networks, wireless nodes are owned by rational and intelligent users. We call nodes selfish if they are owned by independent users and their objective is only to maximize their individual goals. In a broadcasting task for wireless ad hoc networks, a source node sends the same message to all the nodes in the network. Broadcast is useful in route discovery, and paging a particular host or sending an alarm signal. So, it is interesting to study broadcast in which to consider incentives because wireless nodes are belonging to independent, self-interested users. Most of the existing protocols for broadcast assume that nodes voluntarily follow the prescribed protocols without deviation. These protocols will not work if the nodes exhibit selfish behavior. Thus there is a need for truthful computing in this context. We call this problem the truthful broadcast problem for wireless ad hoc networks. Truthful computing refers to a stimulation mechanism to make selfish nodes cooperate with each other. Providing incentives and pricing the transactions are well known approaches to stimulate cooperation. There is a considerable amount of research in the literature on truthful unicast and truthful multicast problems in wireless ad hoc networks with selfish nodes. However, we show that applying the solution approaches of truthful unicast and multicast problems to solving the truthful broadcast problem lead to inefficient solutions. So, in order to solve the truthful broadcast problem, we present a game theoretic framework for this problem and a strategy proof pricing mechanism called Immediate Predecessor Node Pricing Mechanism (IPNPM). The phrase strategy proof here means that truth revelation of cost is a weakly dominant-strategy (in game theoretic terms) for each node. In order to steer our mechanism-design Authorized licensed use limited to: INDIAN INSTITUTE OF SCIENCE. Downloaded on June 12,2010 at 07:21:15 UTC from IEEE Xplore. Restrictions apply.