Transmission Optimization In Large-scale Social Networks With Evolutionary Characteristics

In recent years,the advancement of information technology and the high-speed development of mobile Internet have brought a vigorous development for the social network.Such new development has not only greatly expanded the application scenario of social networks,but also brought out the dynamic evolution and wireless transmission characteristics.Consequently,the large-scale social network with evolutionary characteristics has gradually emerged.Under the new application scenario,network evolution and wireless transmission pose many new challenges to the performance of information transmission.On one hand,network evolution leads to the rapid growth of network scale,as well as the constant changes in user relationships and transmission demands.On the other hand,the information transmission between model devices are based on the wireless channels.However,randomness of wireless channels results in an uncertain transmission process and the unpredictable performance.In a large-scale social network with evolutionary characteristics,a dilemma naturally arises: how to achieve stable transmission performance in dynamic and complex transmission scenarios? To solve this problem,this thesis analyses and optimizes the performance of information transmission in large-scale social networks with evolutionary characteristics.Firstly,in terms of time,this thesis analyzes the transmission efficiency of evolving social networks.Secondly,from the perspective of space,the information transmission capacity of large-scale wireless social networks is studied.Based on the above research foundations,from the perspective of space-time combination,the transmission performance of wireless evolving social networks is further discussed and optimized.The main contents of this thesis are presented as follows:Firstly,this thesis studies the efficiency of information diffusion in evolving social networks.In social networks,the information diffusion is conducted via a spreadingbased diffusion mechanism,where a user simply spreads the information she is interested in to all her friends.This spreading-based diffusion mechanism is blind-guided and results in low diffusion performance.In evolving social networks,the homophilydriven connections between users provide much potential for the improvements of transmission efficiency.In this part,we firstly model the network as a homophilydriven evolving social network,where users cut old connections and try to connect with others who share more same attributes with them.We further compute the transmission precision and recall during the evolution process.Our theoretical results demonstrate that during the evolution process,the users with more same attributes will connect with each other.In this way,users will spontaneously forward the information they are interested in to their neighbors,so that users who receive the information will also be interested in it,and ultimately the transmission precision and recall will be greatly improved.Specifically,when the evolution process converges to the stable state,similar users will be closely connected,and the precision and recall can achieve to the optimum.That is,the users who receive the information are interested in it,and the users who are not interested in the information will not be informed.Secondly,wireless social networks have been widely used in a series of scenes,such as earthquake relief,military battlefield and so on.In a large-scale wireless social network,massive users bring unprecedented transmission load,while the random channel leads to unstable transmission process.To tackle this challenge,we study the network capacity in wireless social networks.We apply the small-world model to characterize the social relationships between users.In detail,each user has four local contacts and several long-range contacts.The transmission probability between users is inversely proportional to the physical distance between them.In order to improve the network performance,we consider that each user is equipped with a single-beam or multi-beam directional antenna.Next,we analyze and optimize the network capacity of wireless networks with directional antennas.For both single-beam and multi-beam directional antennas,the corresponding lower and upper bounds of network capacity are derived.Our theoretical results show that the lower bounds of the network capacity are consistent with upper bounds in the asymptotic performance for both two types of antennas.Moreover,we prove that stronger social relationships and more concentrated antenna beams could significantly improve network performance.Specifically,when the social contact density and the beamwidth of antenna main lobe satisfy certain conditions,the network capacity can achieve the optimal performance.At last,based on the foundations of the above research,from the perspective of space-time combination,we comprehensively consider the complex scenarios in which network evolution and wireless transmission characteristics coexist.To address the conflicts between complex evolution scenarios and high-speed transmission demands,we study the delivery rate in a wireless evolving social network.Considering the cooperative evolution scenario of users and transmission content in a network,we utilize the affiliation network to characterize the evolution process,and derive the power-law distribution of both user degrees and content popularity.In view of the power-law distribution,we use cache technology to improve content delivery performance.Specifically,we formulate the process of solving cache storage strategy into a convex problem,and use the KKT condition to solve this problem.Further,the corresponding content delivery rate is derived.Our theoretical results show that in a wireless evolving social network,when the content popularity is sharply distributed,i.e.,most of the users are interested in a small number of popular files,the delivery rate can achieve the optimum.