Journal 2013 - Incentive Mechanism for Multiuser Cooperative Relaying in Wireless Ad Hoc Networks - A Resource-Exchange Based Approach

Incentive Mechanism
of 19
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Wireless Pers Commun (2013) 73:697–715DOI 10.1007/s11277-013-1211-z Incentive Mechanism for Multiuser Cooperative Relayingin Wireless Ad Hoc Networks: A Resource-ExchangeBased Approach Guopeng Zhang  ·  Kun Yang  ·  Peng Liu  · Xiaolong Feng Published online: 26 May 2013© Springer Science+Business Media New York 2013 Abstract  This paper studies the resource allocation (RA) and the relay selection (RS) prob-lems in cooperative relaying (CR) based multiuser ad hoc networks, and a multiuser cooper-ative game is proposed to stimulate selfish user nodes to participate in the CR. The noveltyof the game scheme lies in that it takes explicit count of that a wireless user can act as adata-source as well as a potential relay for other users. Consider a user has the selfish incen-tive to consume his/her spectrum resource solely to maximize his/her own data-rate and theselection cooperation (SC) rule which restricts relaying for a user to only one relay is explic-itly imposed. To stimulate user nodes to share their energy and spectrum resource efficientlyin the Pareto optimal sense, first, we formulate the RA problem for multiuser CR as a bar-gaining game. By solving the Nash bargaining solution of the game, Pareto optimal RA forcooperative partners can be achieved. Next, to implement the SC-rule imposed RS, a simpleheuristic is proposed with the main method being to maintain the long-term priority fairnessfor cooperative partner selection for each selfish user. The proposed RS with RA (RS-RA)algorithm has a low computational complexity of   O ( K  2 ) , where  K   is the number of usersin a network. Simulation results demonstrate the system efficiency and fairness properties of the proposed bargaining game theoretic RS-RA scheme. G. Zhang ( B )  ·  P. LiuInternet of Things Research Center, China University of Mining and Technology,Xuzhou 221008, Jiangsu, Chinae-mail: Liue-mail: YangSchool of Computer Science and Electronic Engineering (CSEE), University of Essex, Wivenhoe Park,Colchester CO4 3SQ, Essex, UKe-mail: X. FengSchool of Information and Electrical Engineering, China University of Mining and Technology,Xuzhou 221116, Jiangsu, Chinae-mail:  1 3  698 G. Zhang et al. Keywords  Cooperative relaying  ·  Resource allocation  ·  Relay selection  ·  Cooperativebargaining game  ·  Nash bargaining solution  ·  Pareto optimal 1 Introduction Cooperative relaying (CR) [1] allows exploiting the spatial diversity gains inherent in multi- user wireless networks without the need of multiple antennas at each user node. Recently,CR techniques have evolved from the early information theoretic results [2] to a practical application stage. For example, fixed infrastructure based relay nodes (RNs) have emergedin cellular networks [3,4], and multi-hop communication via the RNs can effectively extend base stations (BSs) coverage. In cellular networks, fixed RNs are dedicatedly deployed tohelp wireless users’ uplink/downlink transmissions without having their own data to trans-mit. However, an ad hoc network is composed of multiple distributed user nodes and doesnot rely on any infrastructure [5]. Each user node plays an equal role and has its own datato transmit. It thus can act as a data-source as well as a potential relay for other nodes.In such decentralized network environments, a node has the right to determine whether tocooperate with others according to its own willingness, and it tends to be selfish to consumeits resource solely to optimize its own performance [6]. Therefore, cooperation cannot be taken for granted in such a selfish scenario. The relay selection (RS) and resource allocation(RA) problem become more complicated. It is necessary to develop cooperation stimulatingmechanisms that allow cooperation to emerge in the presence of selfish nodes.Inliteratures,cooperationincentivemechanismsforselfishnodesarebroadlydividedintothe following three categories. ã  Reputation based mechanisms [7–9], in which cooperation behaviors of nodes are mea- sured by other nodes in the same network. Selfish behaviors are then discouraged by thethreat of partial or total refusal of forwarding of any traffic srcinated from these selfishnodes. However, reputation mechanisms always rely on the usage of tamper-proof hard-ware to store and check reputation credit. This strategy may hinder their ability to findwide-spread acceptance. ã  Pricing based mechanisms [10–13], in which relay nodes are usually assumed to have temporarilyidledspectrumandenergyresourcethatarenotfullyutilized.Thereforetheyhave opportunities to sell the idle resource to the nodes that are seeking extra spectrum(via cooperation), and thereby to generate revenue. The revenue can be used by the relaynodes later to reward and encourage other nodes to cooperate. In this research area,Huang [10] and Zhang and Cong [11,12] investigate the problems of how a relay should sell its power and bandwidth to multiple competing source nodes using auction theoryand non-cooperative pricing game theory, respectively. Based on the Stackberg pricinggame theory, Wang [13] studies the scenario in which multiple relays compete with each other in terms of price to gain the highest profit from selling power to a single source.Although pricing mechanisms could stimulate selfish nodes to cooperate, it does notfunction in full-loaded network situations where all nodes in a wireless system alwayshave their own data to transmit, and, thus leads to a situation where some nodes as relayswill have no more resource left for selling to other nodes. ã  Resource-exchangebasedmechanisms[14–16],inwhichonenodecantradeitsresource for the partner node’s CR directly, and, as a result, two cooperative partner nodes can actas a source as well as a potential relay for each other symmetrically. By this way, firstly,the tamper-proof hardware used in reputation mechanisms can be avoided. Secondly, the  1 3  A Resource-Exchange Based Approach 699 full-loaded situation mentioned above can also be dealt with effectively, since resource-exchangemechanismsprovidecooperativeincentivesforselfishusernodesbypromisingto lead to a “win-win” situation, e.g., transmission rate gains for both cooperative partnernodes. However, the authors in [14] and [15] only consider the special case of two-node cooperation in cellular and ad hoc networks, respectively. As the selection coopera-tion (SC) rule which restricts relaying for a source node to only one relay is explicitlyimposed in both the works, the authors do not actually touch the RS problem, whichis an important issue in multiuser networks. Moreover, in [15], the authors also studythe bandwidth allocation in the one-to-many CR network model. However, again, theRS problem is not discussed. It requires the extension of the one-to-many CR model toaccommodate the many-to-many case. In [16], the authors provide a resource-exchangebased non-cooperative game to perform the RA in the paradigm of CR based cognitiveradio networks.Different from [14–16], in this paper, we study the resource-exchange based cooperation stimulatingmechanismforCRbasedmultiuseradhocnetworks,andtheRSandRAproblemsare considered jointly. Specifically, our main contributions include the following: ã  A resource-exchange based cooperative bargaining game is proposed, whose aim is tostimulate a subset of potential relays of a source node to forward the source’s data. In thegame, to stimulate a relay to cooperate, the source should reward the relay by forwardingthe data srcinated from the relay. Assuming each selfish node is willing to cooperateonlyifthetransmissionrateachievedthroughcooperationisnotlowerthanthatachievedwithout the cooperation by consuming the same amount of energy, we can prove thereexists a unique Nash bargaining solution (NBS) for the proposed game. By solving theNBS, all the cooperative partner nodes can achieve rate-gains in comparison with thatachieved through direct transmission. ã  As the algorithm to solve the srcinal game is with a high computational complexity of  O (  L 2 K  ) , where  K   is the total number of the potential relays available to a source and  L  is the number of data symbols that can be transmitted in a frame of a source, the SCrule is applied to the game which could also bring the cooperative partner nodes win-win benefits and the solution method to the SC-rule based game is featured with a lowcomputational complexity of   O ( K  ) . ã  ToimplementtheRSandtheRAjointly,asimpleheuristicRSwithRA(RS-RA)schemeis proposed to maintain the long-term priority fairness for the cooperative partner selec-tion. Based on the SC rule, the method of the RS-RA heuristic is to assign each sourcea different priority at the network initialization stage, and a source with a higher priorityhas privilege to select its cooperative partner. With the priority of each source beingupdated after each RS-RA cycle, the long-term priority fairness for the whole network can be guaranteed.The rest of this paper is organized as follows. In Section II, we propose the investigatedmultiuser CR-based ad hoc network model. In Sect. 3, we describe the resource-exchange based cooperation incentive mechanism. In Sect. 4, we formulate the RA problem for coop- erative nodes as a cooperative bargaining game, and we also prove the existence of a uniqueNBStothegame.InSect.5,asimplifiedSC-ruleimposedmultiuserCRgameisproposedandis solved in a closed form. In Sect. 6, a simple heuristic RS scheme is proposed to maintain the long-term priority fairness for cooperative partner selection in the network. Simulationresults are presented in Sect. 7. Finally, conclusions are drawn in Sect. 8.  1 3  700 G. Zhang et al. Fig. 1  Multiuser CR basedad hoc network  2 System Model Inwirelessadhocnetworks,anend-to-enddata-flowmaycrossseveralhopsinthelinklayer.The rate requested by the flow is checked against the link layer capacity hop-by-hop to find afeasible path. The data-rate that can be achieved by such flow is determined by the capacityof the bottleneck link. As a first step to study the resource-exchange based CR stimulatingmechanism, this paper only focuses on the one-hop CR case. The studied multiuser ad hocnetwork is sketched in Fig. 1. It consists of   K   + 1 one-hop source-to-destination links, fromsource  f  i  to destination  F  i , i  =  1 ,...,  K  , and from source  m  to destination  M  , where  h m ,  M  represents the channel from source  m  to destination  M  , h m ,  f  i  represents the channel fromsource  m  to source  f  i , h  f  i ,  M   represents the channel from source  f  i  to destination  M  , h  f  i , F  i represents the channel from source  f  i  to destination  F  i , h  f  i , m  represents the channel fromsource  m  to source  f  i , and  h m , F  i  represents the channel from source  m  to destination  F  i .Notethatweonlyfocusonthesources’cooperation,i.e.,onlysources { m ,  f  1 ,...,  f  K  } canserve as relays for other sources. The cooperation results for destinations  {  M  ,  F  1 ,...,  F  K  } canbeobtainedinasimilarwaywhichisomittedhere.Consideringthedirectlinkfromsource m  to destination  M   is weak due to the long distance transmission or the line of sight (LoS)being obstructed by a building block, other sources  {  f  1 ,...,  f  K  }  may serve as the relays forsource  m ’s transmission, in view of CR may provide intermediate facility for improve rateperformance. Since all sources are peer in ad hoc network setups, so each potential relay  f  i ( 1  ≤  i  ≤  K  )  also has the right to serve as a source.Throughout the following analysis, we assume the system bandwidth is  W   Hz and theorthogonalchannelsformultipleaccessaresynthesizedbyexistingCRorientedcarriersensemultiple access with collision avoidance (CSMA/CA) protocols, e.g., [17]. We also assumethesourcesarescheduledtotransmitinarapidsuccession,oneaftertheother,eachusingoneschedulingtimeintervalcalledadata-frame.Adata-framecarryingseveraldatasymbolsthenrepresents one basic time-frequency unit for the sequel CR based RA. We consider the worstfull-loaded network situation, where all sources  { m ,  f  1 ,...,  f  K  }  are active simultaneouslyandhaveinfinitebackloggeddatatosend.Thus,thepricingbasedCRstimulatingmechanism[10–13] do not function in such situations since there is no extra idle spectrum resource left forthesourcestolease.Thenoveltyoftheproposedresource-exchangebasedCRstimulatingmechanism is in taking explicitly account of that each cooperative source has its own data totransmit and all the cooperative sources could get performance gains through the CR.Recently, new emerged full-duplex (FD) CR offers great potential for increasing spectral-efficiency than traditional half-duplex (HD) CR [18]. The FD mode can receive and forwardsimultaneously on the same frequency-band, however, it orders that each node should beequipped with at least two isolated receive and transmit antennas which is infeasible for  1 3
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks