A Game Theoretic Approach for Content Distribution over Wireless Networks with Mobileto-Mobile

Transcription

1 22 nd Annual IEEE International Symposium on Personal, Indoor and Mobile Radio Communications A Game Theoretic Approach for Content Distribution over Wireless Networks with Mobileto-Mobile Cooperation Lina Al-Kanj 1, Walid Saad 2, Zaher Dawy 1 1 Electrical and Computer Engineering Department, American University of Beirut, Beirut, Lebanon 2 Electrical and Computer Engineering Department, University of Miami, Coral Gables, FL, USA Toronto, Canada September 13, 2011

2 Outline Introduction Literature Review Motivation Coalition Formation Game Results and Analysis Conclusion 2

3 Introduction Emerging mobile applications require the mobile terminals (MT)s wireless interfaces to be active for long periods which results in high energy consumption. Studies show that the high energy consumption of battery-operated MTs will be one of the main limiting factors for future wireless communication systems. Several heuristic algorithms for Cooperative Content Distribution have been investigated in the literature to decrease the energy consumption in the MTs. 3

4 Introduction Requesting mobile terminals (MTs) interested in a common content MTs connect to a BS/AP to download the content BS-MT: Long Range (LR) Links (unicast transmission) MTs communicate with each other: MT-MT: Short Range (SR) Link (unicast or multicast transmission) Objective : develop distributed content distribution strategies to minimize the total energy consumption in the network 4

5 Literature Review Cooperative content distribution over wireless networks: Performance metrics (energy, throughput, etc.) Long-range distribution strategy Short-range distribution strategy Grouping/clustering of MTs Approach (analytical, simulation, experimental) 5

6 Literature Review Reference Metric LR distribution SR distribution Clustering Approach Fitzek et al. [1] Energy consumption MTs pull parts randomly Multicasting/un icasting All MTs cooperate Analytical & experimental Fitzek et al. [2] Energy consumption AP broadcasts SR retransmission All MTs cooperate Analytical & experimental CHUM [3,4] End user Cost/Energy consumption MTs pull parts randomly Multicast to members of group chumcast server management Analytical & simulated COSMOS [5] Throughput & End user cost MTs pull parts randomly Broadcast Ad hoc mode with a limited number of hops Simulated Popova et al. [6] Throughput & download time BS distributes packets 1by1 to MTs Multicast to members of group BS manages group formation Simulated ICAM [7] Throughput BS delivers content to MTs with good channel quality MT forward to other MTs through ad hoc mode Cooperation within grids determined by BS Polynomial time app. Alg 6

7 Distributed Game Theoretic Approach: Motivation Solution does not rely on a central entity MTs cooperate based on local information We propose a distributed game theoretic approach to form the coalitions and distribute the content The proposed coalition formation algorithm is based on merge-and-split operations that proved to reach a stable solution that is close to the optimal solution. 7

8 Problem Formulation: Parameters Parameters K Number of MTs, M k,, MT (k=1,, K) Channel Conditions S T Data Rates Consumed Energies P e d k d kj Distances between BS-M k, M k - M j h k h kj Channel gains between BS-M k, M k - M j Size of the content in MB R L,k Rate on LR (k=1,, K) in Mbits/sec R S,kj Rate on SR (k,j=1,,k) in Mbits/sec E Rx,L Consumed energy per second while receiving on LR E Rx,S Consumed energy per second while receiving on SR E Tx Consumed energy per second on while transmitting on SR Probability of error 8

9 Problem Formulation: All MTs Cooperate Energy Consumed to receive on LR Energy Consumed to transmit on SR Energy Consumed to receive on SR The payoff of any MT k G The utility of each coalition 9

10 Merge-and-Split: Operations Merge Operation: Merge any set of coalitions *G 1,, G l + if: v l l i=1 G l < i=1 v(g l ) Split Operation: Merge any set of coalitions v l l i=1 G l > i=1 v(g l ) l i=1 G l if: After a number iterations consisting of sequential merge-and-split operations, the network converges to a partition that is merge-and-split proof, i.e., no coalition has any further incentive to join another coalition or to split into smaller coalitions. The convergence of merge-and-split iterations is guaranteed, regardless of the initial partition, as proven in [9]. [9] K. Apt and A. Witzel, A generic approach to coalition formation, in Proc. of the Int. Workshop on Computational Social Choice (COMSOC), (Amsterdam, the Netherlands), Dec

11 Distributed Algorithm: Design A Coalition Formation Game (CFG) for Cooperative Content Distribution 11

12 Results and Analysis: Coalition Formation BS transmit power is higher than MT transmit power The coalition formation game results in MTs energy savings (ES). a)cgf uni/uni (ES 12%). b)cfg uni/multi (ES 25%). Less cooperation takes place if MTs are close to BS More cooperation takes place if MTs are far from BS 12 d)cfg uni/uni (ES 32%). e)cfg uni/multi (ES 48%). Figure 3: Resulting for K = 15 requesting MTs.

13 Results and Analysis: Energy Consumption Multicasting is more energy efficient than unicasting on the SR. Multicasting on SR: the energy savings are up to 60% relative to the noncooperative case and up to 37.5% relative to the grand coalition case, at K = 50. Unicasting on SR: the energy savings are up to 43% relative to the noncooperative case and up to 30% relative to the grand coalition case, at K = Figure 4: Average energy consumption versus K.

14 Results and Analysis: Group Sizes and LR Resources Network composed of large number of small coalition rather than a small number of large coalitions. Multicasting increases the group sizes and decreases the required LR links. The wireless networks resources are significantly saved (up to 72%). 14 Figure 6: Group sizes for K=15 MTs. Figure 7: Required LR links versus K.

15 Results and Analysis: Merge-Split Operations The average number of merge and split operations increases with the number of requesting MTs K. The average number of iterations till convergence increases with the number of requesting MTs K. 15 Figure 6: Number of iterations versus K. Figure 7: Average number of merge and split operations per MT versus K.

16 Conclusion We have presented a novel game-theoretic approach based on coalition formation games for content distribution over wireless networks. Within each coalition, a mobile terminal, selected as a coalition head, receives the content from the base station and transmits it to the other mobiles within the same coalition. Results show that the energy consumption decreases up to 60% relative to the non-cooperative case and up to 37.5% relative to the case where all MTs form a grand coalition. The required number of the long-range wireless links can be decreased by up to 72% which maps to significant network throughput gains. 16

17 Thank You for your attention Questions? 17

18 References [1] F. Fitzek and M. Katz, Cooperation in Wireless Networks: Principles and Applications. Springer, ISBN , [2] Q. Zhang, F. Fitzek and V. B. Iversen, Design and Performance Evaluation of Cooperative Retransmission Scheme for Reliable Multicast Services in Cellular Controlled P2P Networks, Proc. of IEEE Int. Symp. on Personal, Indoor and Mobile Radio Communications (PIMRC), September [3] S. Kang and M. Mutka, A Mobile Peer-to-Peer Approach for Multimedia Content Sharing Using 3G/WLAN Dual Mode Channels, Journal of Wireless Communications and Mobile Computing, vol. 5, no. 6, pp , September [4] D. Zhu and M. Mutka, Cooperation among peers in an ad hoc network to support an energy efficient IM service, Journal of Pervasive and Mobile Computing, vol. 4, no. 3, pp , June [5] M. F. Leung and S.-H. Gary Chan, Broadcast-Based Peer-to-Peer Collaborative Video Streaming Among Mobiles, IEEE Transactions on Broadcasting, vol. 53, no. 1, pp , March [6] L. Popova, T. Herpel, W. Gerstacker, and W. Koch, Cooperative Mobile-to-Mobile File Dissemination in Cellular Networks within a Unified Radio Interface, Int. Journal of Computer and Telecommunications Networking, vol. 52, no. 6, pp , January [7] R. Bhatia, L. E. Li, H. Luo and R. Ramjee, ICAM: Integrated Cellular and Ad Hoc Multicast, IEEE Transactions on Mobile Computing, vol. 5, no. 8, pp , August