A. James Clark School of Engineering

Permanent URI for this communityhttp://hdl.handle.net/1903/1654

The collections in this community comprise faculty research works, as well as graduate theses and dissertations.

Browse

Subject: search.filters.subject.social network

Search Results

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Joint Optimization for Social Content Delivery in Wireless Networks
    (2016) Weng, Xiangnan; Baras, John S; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)
    Over the last decade, success of social networks has significantly reshaped how people consume information. Recommendation of contents based on user profiles is well-received. However, as users become dominantly mobile, little is done to consider the impacts of the wireless environment, especially the capacity constraints and changing channel. In this dissertation, we investigate a centralized wireless content delivery system, aiming to optimize overall user experience given the capacity constraints of the wireless networks, by deciding what contents to deliver, when and how. We propose a scheduling framework that incorporates content-based reward and deliverability. Our approach utilizes the broadcast nature of wireless communication and social nature of content, by multicasting and precaching. Results indicate this novel joint optimization approach outperforms existing layered systems that separate recommendation and delivery, especially when the wireless network is operating at maximum capacity. Utilizing limited number of transmission modes, we significantly reduce the complexity of the optimization. We also introduce the design of a hybrid system to handle transmissions for both system recommended contents ('push') and active user requests ('pull'). Further, we extend the joint optimization framework to the wireless infrastructure with multiple base stations. The problem becomes much harder in that there are many more system configurations, including but not limited to power allocation and how resources are shared among the base stations ('out-of-band' in which base stations transmit with dedicated spectrum resources, thus no interference; and 'in-band' in which they share the spectrum and need to mitigate interference). We propose a scalable two-phase scheduling framework: 1) each base station obtains delivery decisions and resource allocation individually; 2) the system consolidates the decisions and allocations, reducing redundant transmissions. Additionally, if the social network applications could provide the predictions of how the social contents disseminate, the wireless networks could schedule the transmissions accordingly and significantly improve the dissemination performance by reducing the delivery delay. We propose a novel method utilizing: 1) hybrid systems to handle active disseminating requests; and 2) predictions of dissemination dynamics from the social network applications. This method could mitigate the performance degradation for content dissemination due to wireless delivery delay. Results indicate that our proposed system design is both efficient and easy to implement.
  • Thumbnail Image
    Item
    Diffusion Dynamics in Interconnected Communities
    (2015) Wei, Xiaoya; Abed, Eyad H.; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)
    In this dissertation, multi-community-based Susceptible-Infected-Recovered (SIR) and Susceptible-Infected-Susceptible (SIS) models of infection/innovation diffusion are introduced for heterogeneous social networks in which agents are viewed as belonging to one of a finite number of communities. Agents are assumed to have well-mixed interactions within and between communities. The communities are connected through a backbone graph which defines an overall network structure for the models. The models are used to determine conditions for outbreak of an initial infection. The role of the strengths of the connections between communities in the development of an outbreak as well as long-term behavior of the diffusion is also studied. Percolation theory is brought to bear on these questions as an independent approach separate from the main dynamic multi-community modeling approach of the dissertation. Results obtained using both approaches are compared and found to be in agreement in the limit of infinitely large populations in all communities. Based on the proposed models, three classes of marketing problems are formulated and studied: referral marketing, seeding marketing and dynamic marketing. It is found that referral marketing can be optimized relatively easily because the associated optimization problem can be formulated as a convex optimization. Also, both seeding marketing and dynamic marketing are shown to enjoy a useful property, namely ``continuous monotone submodularity." Based on this property, a greedy heuristic is proposed which yields solutions with approximation ratio no less than 1-1/e. Also, dynamic marketing for SIS models is reformulated into an equivalent convex optimization to obtain an optimal solution. Both cost minimization and trade-off of cost and profit are analyzed. Next, the proposed modeling framework is applied to study competition of multiple companies in marketing of similar products. Marketing of two classes of such products are considered, namely marketing of durable consumer goods (DCG) and fast-moving consumer goods (FMCG). It is shown that an epsilon-equilibrium exists in the DCG marketing game and a pure Nash equilibrium exists in the FMCG marketing game. The Price of Anarchy (PoA) in both marketing games is found to be bounded by 2. Also, it is shown that any two Nash equilibria for the FMCG marketing game agree almost everywhere, and a distributed algorithm converging to the Nash equilibrium is designed for the FMCG marketing game. Finally, a preliminary investigation is carried out to explore possible concepts of network centrality for diffusions. In a diffusion process, the centrality of a node should reflect the influence that the node has on the network over time. Among the preliminary observations in this work, it is found that when an infection does not break out, diffusion centrality is closely related to Katz centrality; when an infection does break out, diffusion centrality is closely related to eigenvector centrality.
  • Thumbnail Image
    Item
    Behavior Modeling and Forensics for Multimedia Social Networks
    (2009) Lin, Wan-Yi; Liu, K.J. Ray; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)
    Within the past decades, the explosive combination of multimedia signal processing, communications and networking technologies has facilitated the sharing of digital multimedia data and enabled pervasive digital media distribution over all kinds of networks. People involved in the sharing and distribution of multimedia contents form \emph{multimedia social networks} in which users share and exchange multimedia content, as well as other resources. Users in a multimedia social network have different objectives and influence each other's decision and performance. It is of ample importance to understand how users interact with and respond to each other and analyze the impact of human factors on multimedia systems. This thesis illustrates various aspects of issues and problems in multimedia social networks via two case studies of human behavior in multimedia fingerprinting and peer-to-peer live streaming. Since media security and content protection is a major issue in current multimedia systems, this thesis first studies the user dynamics of multimedia fingerprinting social networks. We investigate the side information which improves the traitor-tracing performance and provide the optimal strategies for both users (fingerprint detector and the colluders) in the multimedia fingerprinting social network. Furthermore, before a collusion being successfully mounted, the colluders must be stimulated to cooperate with each other and all colluders have to agree on the attack strategy. Therefore, not all types of collusion are possible. We reduce the possible collusion set by analyzing the incentives and bargaining behavior among colluders. We show that the optimal strategies designed based on human behavior can provide more information to the fingerprint detector and effectively improve the collusion resistance. The second part of this thesis focuses on understanding modelling and analyzing user dynamics for users in various types of peer-to-peer live streaming social networks. We stimulate user cooperation by designing the optimal, cheat-proof, and attack-resistant strategies for peer-to-peer live streaming social networks over Internet as well as wireless networks. Also, as more and more smart-phone users subscribe to the live-streaming service, a reasonable market price has to be set to prevent the users from reselling the live video. We start from analyzing the equilibrium between the users who want to resell the video and the potential buyers to provide the optimal price for the content owner.