基于博弈论的云用户行为分析预测模型研究与分析

发布时间：2018-04-08 09:56

本文选题：云计算　切入点：用户行为　出处：《东华理工大学》2016年硕士论文

【摘要】：随着云计算技术的蓬勃发展,云计算的新理论、新应用正慢慢受到人们的认可与喜爱。在云计算技术逐步走向主流的过程中,其新产品也由少到多、逐渐侵吞传统网络技术产品的市场。然而受限于云计算本身的技术条件和工作模式,其安全问题也渐渐显露出来,受到学者和大众的广泛关注。如何更好解决已有的缺陷和未知的隐患,为云计算技术更好更快的普及保驾护航等,都是当下亟待解决的问题。本文以云用户行为为研究对象,运用经济学中的博弈理论对其进行研究。将云用户和云服务商看作对立的两方博弈者,首先,运用不完全信息静态博弈理论,建立了一种基于不完全信息静态博弈的云用户行为分析预测模型,分析得到模型中存在的均衡,再通过分析不同损益项的变化给云用户行为带来的影响,得出合理的安全管理方案;其次针对不完全信息静态博弈分析条件与现实情况存在差距,运用不完全信息动态博弈理论,由云用户的行动作为计算依据,得出云服务商对云用户类型的评估,再进行重复博弈,从云用户的历史行动使得云服务商的评估得到更新,对云用户的类型的判断也更为准确,从而达到预测效果;再次,针对传统博弈理论分析条件的欠缺,运用演化博弈理论,将云用户和云服务商看作对立的两方博弈者种群,用复制动态方程描绘参与人行为策略选取趋势,用演化稳定策略刻画动态局面的收敛方向,用雅可比矩阵验证平衡点的局部稳定性,分析讨论了不同损益条件对云用户行为的影响,在此基础上给出合理的安全方案,供管理人员参考。本文也对上述3个内容进行了实验分析,验证了模型的理论,以及预测结果的正确性。对云用户行为进行分析预测可以在云用户没有进行任何破坏行为之前提前控制用户的行为,达到从较早环节抑制异常行为发生的作用,使安全防护具有主动性和行为预见性,运用博弈理论对用户行为进行可信研究,为管理人员安全策略的制定提供了更多选择,也为安全管理提供了一种新思路。
[Abstract]:With the rapid development of cloud computing technology, the new theory of cloud computing, new applications are gradually recognized and loved by people.In the process of cloud computing technology gradually moving to the mainstream, its new products from less to more, gradually embezzled the market of traditional network technology products.However, limited by the technical conditions and working mode of cloud computing itself, the security problem of cloud computing has gradually emerged, which has been widely concerned by scholars and the public.How to better solve the existing defects and unknown hidden dangers, for cloud computing technology better and faster to protect the popularity of such issues are urgent need to be solved.In this paper, cloud user behavior is studied by using game theory in economics.Cloud users and cloud service providers are regarded as opposing game players. Firstly, using the theory of incomplete information static game, a cloud user behavior analysis and prediction model based on incomplete information static game is established.By analyzing the influence of different profit and loss items on cloud users' behavior, the reasonable security management scheme is obtained. Secondly, there is a gap between the static game analysis conditions of incomplete information and the actual situation.Based on the dynamic game theory of incomplete information, the cloud service providers' evaluation of cloud users' types is obtained, and then repeated games are used to update the cloud service providers' evaluation.The types of cloud users are also more accurate to achieve the prediction effect. Thirdly, the cloud users and cloud service providers are regarded as opposing two-party gamblers population by using evolutionary game theory in view of the lack of analysis conditions of traditional game theory.Using the replicating dynamic equation to describe the behavior strategy selection trend of the participants, using the evolutionary stability strategy to describe the convergence direction of the dynamic situation, and using Jacobian matrix to verify the local stability of the equilibrium point.This paper analyzes and discusses the influence of different profit and loss conditions on the behavior of cloud users, and gives a reasonable security scheme for managers' reference.This paper also carries on the experimental analysis to the above three contents, verifies the theory of the model and the correctness of the prediction results.The analysis and prediction of cloud user's behavior can control the cloud user's behavior in advance before the cloud user has any destructive behavior, so as to restrain the abnormal behavior from the earlier link, so that the safety protection has initiative and behavior foresight.Using game theory to conduct credible research on user behavior provides more choices for the formulation of safety policy for managers and a new way of thinking for security management.
【学位授予单位】：东华理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP393.08

【参考文献】