Research On The Optimization Of Network Resource Allocation Based On Game Theory And Mechanism Design

The Internet has become the information infrastructure of human society.With the exploding demands on Internet contents,the Internet is facing a huge pressure brought by massive traffic.It becomes an important common concern in academia and industry to efficiently allocate network resources.The Internet is a complicated computer system that involves many economic issues.Economic factors have become important considerations and economic research models become effective tools for network resource allocation.The interactions among multiple participants under complicated scenarios and the allocation of new resources in mobile communication need to be further studied.This thesis studies the optimization of network resource allocation based on game theory and mechanism design theory.Specifically,it studies the game based analysis of Internet markets and designs efficient network resource allocation mechanisms.The major contents and contributions include:1.To deal with the unbalanced profit distribution caused by P2 P technology in Internet content market,two interrelated game models,including the strategy-chosen game and the price-decision game,are set up to analyse the interactions among network operators,content providers and users.A theorem is proposed to solve the price-decision game,according to the result of which we analyse the possible states of the market and the migrations among the states.Further,numerical analyses show that there exists no win-win situation under non-cooperative games and the bandwidth resources are underutilized;2.A framework for the cooperation between network operators and P2 P content providers is proposed.Based on the models of revenue,cost and profit,a solution of the cooperative game is designed based on Nash bargaining theory.Fair profit distribution mechanisms are proposed for both inter-coalition and inner-coalition profit sharing.Potential operator competitions under the cooperation framework are further analyzed.Numerical analyses show that under cooperation and the profit distribution mechanisms,the profits of network operator and that of content provider raise by 145% and 90%,respectively,through efficient resource utilization;3.A framework for the integrated utilization of heterogeneous network resources(cellular and WLAN)is proposed,where users can reveal their valuations on WLAN resource via simple bids.An auction mechanism for alternative resources is designed,and is proved to be incentive-compatible and individually-rational under both complete information and incomplete information,respectively.From the perspectives of mobile operators and the society,profit and social welfare are optimized respectively.Real trace based simulations are carried out and the evaluation results show that the operator profit raises by 25%-40%,and the social welfare raises by 47%.Under incomplete information,the profit and social welfare achieve 80% and 85% of theoretical optimal values;4.To solve the matching of supplies and demands of mobile network edge storage,a region model is built for the market,base on which we design a pre-match algorithm to turn the original supplies and demands into multi-type divisible goods and bundle demands.To achieve the goal of social welfare optimization,a double auction for multi-type resources is proposed and proved to satisfy incentive-compatible,individually rational and weakly budget balanced.To achieve the above three properties,the mechanism has to sacrifice some efficiency.Real trace driven simulations show that the social welfare achieves 72%-91% of the theoretical optimal level.