In this thesis we analyze the performance of communication networks using game theoretic approaches. The thesis is in two parts. The first part studies the performance of Ad hoc, cellular and transportation networks taking into consideration spatial effects. The second part studies economic issues in the communication networks related to the 'net neutrality' regulations. Here we study price competition and revenue sharing mechanisms between the network service providers. In the first part, we use Medium Access Control (MAC) game and Jamming game models to study the performance of a Mobile Ad hoc NETwork (MANET), and routing games to study the performance of a transportation network. In the cellular networks, we study the effect of reducing the spatial density of base stations on the amount of radiation to human body (green networking). We use tools from stochastic geometry to model spatial characters. We begin with the study of MAC game in a MANET where the nodes are noncooperative, and their locations form a Poisson point process. The nodes use Aloha protocol to access the channel, and choose their Medium Access Probability (MAP) such that it optimizes their utility. The utility of each node is defined as weighted difference between a performance metric (expected goodput or expected delay) and transmission costs. We first consider a scenario in which nodes can be priced for channel access. We show that by using a simple (linear) pricing mechanism and setting the price 'appropriately', the selfish behavior of the nodes can be used to achieve the same performance as social optima at equilibrium. In the case where channel access is free, we consider transmission energy costs and analyze the Price of Anarchy (PoA). For the game with goodput based utility, we show that the PoA is infinite at the price that achieves the global optimal goodput. We then study the performance of MANET in the presence of a Jammer while all the nodes cooperate by using a MAP that is assigned by a network an Operator. The objective of the Jammer is to degrade the spatial performance of the MANET by causing interference, whereas the objective of the Operator is to set a MAP to optimize it. We model the interaction between the Jammer and the Operator taking into account the transmission energy costs. We transform the resulting non zero sum game into a zero sum game by constructing an anti-potential, and show that for a given Jammer's transmission cost, if the nodes transmit at a power higher than certain threshold, then the Jammer has no incentive to jam, i.e., the nodes can operate as if there is no Jammer. Next, we consider cellular networks. We study energy saving by switching off a fraction of the base stations (BSs). This saving comes at some cost: the coverage is reduced, and the uplink transmission power of mobiles may increase. This implies exposure of the human body to stronger electromagnetic fields. We quantify this through the deactivation of base stations under the assumptions that the random location of base stations and mobiles form a Poisson processes. We consider the case of sparse network where the interference is negligible, and the case where it is non negligible. We observe that when the mobiles have no power constraints, unlike in the case of negligible interference, switching off base stations reduces the uplink power. Finally, we study a dynamic routing game by adding temporal dimension to the classical routing problem. We consider a scenario where each user have to ship a fixed demand on a shared link. The users can delay shipping their demand, but need to ship within T days, and each time they delay, a delay cost is incurred. We study the effect of delay costs on the user strategies by translating the time dimension into space dimension. Considering both atomic and non-atomic game models and polynomial congestion costs, we show that there exists a threshold time T T before which all users ship their demand at equilibrium, and during this period total demand on the link is decreasing in time. Further, we extend the analysis to the case when the demand of each user arrives randomly. In the second part of the thesis, we study economics aspects in communication networks. Representatives of several Internet access providers (ISPs) have expressed their wish to see a substantial change in the pricing policies in the Internet. In particular, they would like to see content providers (CPs) pay for use of the network, given the large amount of resources they use. This would be in clear violation of the "network neutrality" principle that had characterized the development of the wireline Internet. Our first goal here is to propose and study possible ways of implementing payments in a nonneutral network. We introduce a model that includes the users' behavior, the utilities of the ISP and of the CPs, and the monetary flow that involves the users, an ISP and CPs, and CP's revenues from advertisements. We consider various game models and obtain the resulting equilibrium prices. We show that if a regulator decides the amount of money that CPs pays to the ISP, rather it is decided by the ISP or the CPs, then it results in a favorable situation for all the players. Thus, we demonstrate the necessary for regulation of the monetary transactions between the ISP and the CPs in the nonneutral regime. The mechanism we propose for monetary interaction is based on the Nash bargaining solution. One of the central issues in the debate on network neutrality has been whether one should allow or prevent preferential treatment by an ISP. This raised the question of whether to allow an ISP to have exclusive agreement with a content provider (CP). We study the impact of exclusive contracts between a CP and an ISP in a nonneutral network. We consider a simple case of collusion where a CP and an ISP aim to maximize their sum of revenues, and show that such a collusion may not always be beneficial. We derive an explicit condition in terms of the advertisement revenues of the CPs which tells when a collusion is profitable to the colluding entities. Finally, we consider discrimination in the opposite direction, i.e., CP discriminating the ISPs. We derive models for studying the impact that the CP can have on the utilities of the ISPs by favoring one of them through exclusively revealing its private information on the demand. For a special case of linear demand function, we show that such preferential treatment always benefits the ISPs. However, if the CP charges the ISPs for providing the private information, then the ISP obtaining preferential treatment may not always gain. We then propose mechanisms based on weighted proportional fairness for deciding the payments between the CP and the ISP and compare them by proposing a new metric termed Price of Partial Bargaining.
| Identifer | oai:union.ndltd.org:CCSD/oai:tel.archives-ouvertes.fr:tel-00937510 |
| Date | 06 November 2013 |
| Creators | Hanawal, Manjesh Kumar |
| Publisher | Université d'Avignon |
| Source Sets | CCSD theses-EN-ligne, France |
| Language | French |
| Detected Language | English |
| Type | PhD thesis |
Page generated in 0.0034 seconds