基于博弈论和机制设计的网络资源优化配置研究

发布时间：2018-03-21 08:28

本文选题：网络资源配置　切入点：博弈论　出处：《清华大学》2016年博士论文　论文类型：学位论文

【摘要】：互联网已经成为人类社会的信息基础设施。随着互联网内容需求的飞速增加,互联网面临巨大的流量压力,如何实现有限网络资源的高效配置,成为学术界和工业界共同关心的重要问题。互联网是复杂的计算机系统,除技术因素外还涉及经济因素,在研究网络资源配置的过程中,经济因素成为重要的考虑角度,经济模型是研究网络资源配置的有效工具。复杂场景下的多方参与者互动关系、移动场景下新型资源的高效配置仍有待进一步研究。研究了基于博弈论和机制设计的网络资源优化配置,通过博弈分析和机制设计理论研究了互联网演化中的多方参与者互动和新型资源高效分配。主要研究内容和贡献包括:1.针对P2P技术导致的互联网内容市场利益分配不均,建立了两个相互关联的博弈模型:策略选择博弈和价格决策博弈,对网络运营商、内容提供商和用户之间的互动关系进行分析。针对价格决策博弈,给出了博弈均衡的解析解;基于价格决策博弈的均衡解,分析了策略选择博弈中市场可能存在的状态,以及状态之间的转化关系和转化条件。进一步对非合作博弈下的市场状态进行了数值分析,结果显示,最终不存在共赢结果,带宽资源无法得到高效利用。2.提出了网络运营商和P2P内容提供商的合作框架,在收入、成本和利润建模的基础上进行合作博弈分析,给出了网络运营商与内容提供商之间的纳什讨价还价解,设计了具有公平性的联盟间和联盟内利润分配机制,并对这种合作下网络运营商之间的竞争进行了分析。数值分析显示,在合作和利润分配机制下,网络运营商和P2P内容提供商通过提高资源配置效率,最终利润分别提升了145%和90%。3.设计了移动运营商的异构资源(移动蜂窝网和无线局域网)综合高效利用框架,用户通过简单投标披露对资源的估值,在此基础上,设计了可替代资源拍卖机制,分别在完全信息和不完全信息下实现了拍卖机制的激励相容1?和个体理性,给出了相关证明;并分别对移动运营商利润和社会福利进行优化。实验结果显示,移动运营商利润提升了25%-40%,社会福利提高了47%。在不完全信息下,移动运营商利润和社会福利分别达到了最优目标的80%和85%。4.针对移动网络边缘存储的供需匹配,提出了供需市场的区域模型,在此基础上设计了供需预匹配算法,将初始供需信息转化为多类型可分割商品的供应和多类型商品的组合需求。在优化社会福利的目标下,设计了多类型资源的双向拍卖机制,证明了此机制满足激励相容、个体理性和弱预算平衡。为了满足以上三个特性,将不可避免地损失部分效率。实验结果显示,边缘存储交易机制下的效率达到了理论最优值的72%-91%。
[Abstract]:The Internet has become the information infrastructure of human society. With the rapid increase of Internet content demand, the Internet is facing enormous traffic pressure, how to achieve the efficient allocation of limited network resources. The Internet is a complex computer system, which involves not only technical factors, but also economic factors. In the process of studying the allocation of network resources, economic factors become an important angle of consideration. The economic model is an effective tool for studying the allocation of network resources. The efficient allocation of new resources in mobile scene still needs to be further studied. The optimal allocation of network resources based on game theory and mechanism design is studied. Through game analysis and mechanism design theory, this paper studies the interaction of multiple participants and the efficient allocation of new resources in the evolution of the Internet. The main research contents and contributions include: 1. Aiming at the uneven distribution of the market benefits of Internet content caused by P2P technology, Two interrelated game models are established: strategy selection game and price decision game. The interactive relationship among network operators, content providers and users is analyzed, and the analytical solution of game equilibrium is given for price decision game. Based on the equilibrium solution of the price decision game, this paper analyzes the possible market states in the strategy selection game, the transformation relations and the transformation conditions between the states, and makes a numerical analysis of the market state in the non-cooperative game. The results show that there is no win-win result in the end, and bandwidth resources can not be utilized efficiently. 2. The cooperation framework between network operators and P2P content providers is proposed, and the cooperative game analysis is carried out on the basis of revenue, cost and profit modeling. The Nash bargaining solution between network operators and content providers is given, and a fair profit distribution mechanism between and within coalitions is designed. Numerical analysis shows that under the cooperation and profit distribution mechanism, network operators and P2P content providers improve the efficiency of resource allocation. The final profit is increased by 145% and 90.3. based on the framework of comprehensive and efficient utilization of heterogeneous resources (mobile cellular network and wireless local area network) for mobile operators, users disclose the valuation of resources through simple bidding. The alternative resource auction mechanism is designed and the incentive compatibility of auction mechanism is realized under complete and incomplete information respectively. The results show that the profits of mobile operators have been increased by 25% -40%, and social welfare has increased by 47%. Under incomplete information, the results of the experiments show that the profits of mobile operators have increased by 25% to 40%, and social welfare has increased by 47%. The profit and social welfare of mobile operators reach the optimal goal of 80% and 85.4. aiming at the supply and demand matching of mobile network edge storage, a regional model of supply and demand market is proposed, and a supply and demand pre-matching algorithm is designed on this basis. The initial supply and demand information is transformed into the supply of multi-type divisible goods and the combined demand of multi-type commodities. Under the goal of optimizing social welfare, a two-way auction mechanism of multi-type resources is designed, which proves that the mechanism satisfies the incentive compatibility. Individual rationality and weak budget balance. In order to satisfy the above three characteristics, partial efficiency will inevitably be lost. The experimental results show that the efficiency under the edge storage transaction mechanism reaches 72-91 of the theoretical optimal value.
【学位授予单位】：清华大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：F49;TP393.0

【相似文献】