Research Of Resource Allocation In Multimedia Cloud Based On User Behavior Characteristics

With the advent of the Internet,as well as the rapidly growing popularity of intelligent devices,multimedia services become more and more mainstream and have been widespread attention and application based on video and audio,especially providing live video.However,users are not satisfied to get live video service through desktop computers and watching live by intelligent devices gradually becomes the mainstream.Intelligent devices generally have weak physical properties,limited life and network bandwidth,so that we can't obtain a satisfactory service experience.At the same time,many of the same behavioral characteristics of users have the same demand for multimedia services,so they need to pay for the same multimedia services respectively,resulting in high service costs,underutilized services and other issues.Besides,since the sensitivity of multimedia services on service quality,they have higher requirements for the network delay and computing power of equipment.Therefore,reasonable use and allocation resources of multimedia cloud are very important.In this article,the multimedia cloud system structure was presented based on user behavior characteristics to solve the above problems,users who come close to the location and have the same need of multimedia services are formed into user groups,which constitute a multimedia cloud architecture based on user behavior characteristics.Users within the same group do not have to get all the services from the center multimedia cloud.Users(referred to herein as service users)are chosen who have high utility value to obtain service from the center of multimedia cloud,and then they share it with other users(referred to herein as non-service user).But there are some problems in this multimedia cloud architecture based on user behavior characteristics.First,the stability and the ability to distribution of the service users influence the service experience across user groups,thus the appropriate evaluation standard is required.Secondly,the number of service users is essential,too many service users will result in the overall cost and high customer service costs.Fewer service users will increase the service time cost,resulting in network congestion within the group,greatly reducing non-service users experience.Thirdly,taking into account the user's blindness and selfishness,if you can't balance the bandwidth between service users and non-service users,the bandwidth utilization will be reduced and seriously affect the service experience of the whole system.In view of the above problems,this article is divided into the following three aspects1.For the issue of service user choice in same user group in multimedia cloud architectures,in this article,we consider the user's physical performance and activity,analyze the selfish behavior of users,introduce the credibility of the mechanism and propose user evaluation strategy based on integrated utility value.Users are sorted based on integrated utility in same user group,and the integrated utility value higher,the more suitable to become service users.In addition,for the number of service users,we consider minimize distribution time and minimize cost to solve suitable number of service users.Since this is a multi-objective optimized problem,we propose the algorithm based on extreme disturbed particle swarm for solving the number of service users,in order to achieve the optimization of distribution time and cost.2.For the problem of uneven bandwidth resources distribution in the same user group in multimedia cloud architecture,in this article,we proposed bandwidth resource allocation optimization strategy based on game theory.We take into account blindness and selfishness of non-service users when they choose the service user,so the evolutionary game theory is introduced in the method.In order to obtain greater utility,non-service users change their strategy by imitation,learn and evolution until all non-service users cannot achieve more effectiveness by this strategy,then the evolutionary receives equilibrium in the end.Service users will obtain greater benefits by changing the pricing and sharing bandwidth size to attract non-service users to connect.In this article,we introduce Stackelberg game to analysis competitive behavior of service users and set bandwidth size and pricing,in order to achieve the greatest profits of Nash equilibrium.When pricing strategy and shared bandwidth size of the service user does not change as well as non-service user no longer change the selection policy,user group can achieve the maximum bandwidth allocation of overall effectiveness.3.In this article,we conduct the experiments for the proposed algorithms.First,in order to determine the size of the algorithm parameters,we perform multiple sets of comparative experiments to get maximum performance parameter value algorithm.Then the algorithm based on extreme disturbance particle swarm optimization of choosing service users,weighted extreme algorithm and genetic algorithms are compared,and the collaborative content distribution mechanism and other distribution mechanism are compared at the time overhead and the overhead cost.Finally,game is verified between service users by experiments and the game between non-service users also has been experimental verification,so the balanced bandwidth allocation result is obtained.The result shows that the optimization method has a significant effect for the resources allocation in multimedia cloud in this article.