【作者】 张裕；

【导师】 郑正奇；

【作者基本信息】 华东师范大学 ， 通信与信息系统， 2013， 博士

【摘要】 随着移动通信与信息技术的快速发展,无线通信领域产生了各种无线接入技术以满足人们的生活工作需要。不同的无线网络在接入技术、终端、业务等方面都具有异构性,这样对异构融合网络的可靠性、稳定性、可扩展性提出了挑战。面对多用户,多种业务,高速率的业务需求以及复杂的异构无线环境,为了有效利用异构融合网络中的无线资源,迫切需要高效动态的无线资源管理机制保证整个融合系统具有良好的性能,所以对无线资源管理机制的研究具有重要的意义。本文对异构网络无线资源管理中的网络选择,垂直切换、呼叫接纳控制以及运营商定价与资源分配四个方面的问题进行了探讨,提出了相应的优化机制,并通过建立数学模型与仿真实验及其结果分析证实了算法的有效性,主要研究工作及创新点如下：针对异构无线网络中的网络选择问题从两个方面提出了优化机制,第一个方面对决策因素权重的确定方法进行优化,提出了用可拓模糊层次分析法来获得权重,该方法应用可拓理论计算得到每个决策因素的隶属度,体现了网络间的客观差异程度,然后结合业务类型得到每个决策因素的业务隶属度,再应用模糊层次分析法获得每个决策因素的综合权重,该机制克服了层次分析法中通过人为打分获得权重的主观性问题。第二个方面提出了一种简单有效的优化方法来兼顾用户偏好与运营商偏好,先应用简单加权法分别计算用户偏好与运营商偏好下的各候选网络的性能函数值,并可以根据网络的实际情况动态地设置相应的加权系数,将两种偏好下的各候选网络的性能值相加,具有最大总性能值的网络为最优网络。仿真结果表明提出的网络选择机制在同时考虑了用户偏好与运营商偏好的前提下,有效地平衡了各网络间的负载,提高了整个系统的资源利用率。针对异构无线网络中的垂直切换问题,提出了一种优化方法。将异构无线网络垂直切换决策过程中的切换触发,目标网络选择以及切换执行三个方面作为有机整体进行研究。考虑以SINR作为切换触发判决因素,提出了一种将灰度预测模型与BP神经网络相结合的组合模型预测SINR的方法,以提高SINR的预测精度及切换触发机制的可靠性。接着构建了一个代价函数,能够结合网络实际条件,业务QoS要求以及用户偏好来选择目标网络。仿真结果表明,所提出的垂直切换机制能够提高系统吞吐量,减少垂直切换次数以及降低系统丢话率。针对异构无线网络中的呼叫接纳控制问题,提出了一个最优联合呼叫接纳控制机制,综合考虑了无线网络覆盖范围内不同区域用户业务所能达到的峰值速率及不同的业务类型两个方面的因素,将联合呼叫控制机制建模为一个半马尔可夫决策模型,基于相应的状态空间、行为空间、状态转移概率,并根据业务QoS要求及网络条件提出了一个基于业务类型的接纳效用函数对接纳收益进行评估。在保证业务接纳成功率以及低阻塞率的基础之上,通过最优控制策略使平均接纳效用达到最大值。仿真结果表明所建立的接纳效用函数能够有效反映网络条件的变化对接纳控制的影响,满足业务的QoS要求,获得较高的平均接纳效用,较大的系统吞吐量,同时保证了较低的业务阻塞率。基于存在多个运营商与多个用户的异构无线网络环境,提出了基于博弈论的分布式定价与资源分配机制。首先,研究了运营商间的竞争定价行为,通过最优反应函数获得运营商间的非合作博弈纳什均衡价格。该均衡价格使各个运营商的收益达到最大值。接着,考虑了运营商间的合作博弈来确定最优价格,并提出了一个总收益在运营商间分配的方法,比较和分析了非合作博弈与合作博弈定价对运营商收益的影响。最后,为了体现在运营商定价与资源分配过程中用户行为与运营商行为的相互影响,将运营商定价与用户带宽需求之间的竞争关系建模成一个两阶段的Stackelberg博弈模型,设计了一个考虑用户效用,用户花费,时延花费的用户收益函数,并应用一种只需要局部信息的分布式迭代算法获得了该博弈模型的子博弈完美均衡。最终运营商收益与用户收益均达到了最大值。更多还原

【Abstract】 With the rapid development of mobile communication and information technology, various emerging wireless access technologies out of the area of wireless communication are to satisfy people’s needs of work and life. Different wireless technologies are heterogeneous in terms of access technology, terminal and service, which brings challenges to the reliability, stability and expansibility of heterogeneous integrated network. Facing multi-users, multi-services, high-speed service needs and complex heterogeneous wireless environment, in order to effectively use the wireless resources in heterogeneous network, efficient dynamic management mechanism of wireless resources is highly needed to guarantee fine performance of the whole integrated system. Therefore, the research of wireless resources management mechanism is very significant.This paper researches on network selection, vertical handoff, call admission control and provider pricing for wireless resource management in heterogeneous wireless networks. In the same time, it proposes corresponding optimal schemes that are veritified to be valid through mathematics model and simulation results. Specifically, This paper main contributions can be included in the follows:In the first place, an optimal mechanismfrom two aspects is proposed on how to make a network selection in heterogeneous network, one is to optimize the determination method of decision-making factor weights, and using fuzzy Analytic Hierarchy Process (AHP) and extension theory to obtain weights is raised, by which the degree of membership of each decision-making factor could be obtained through the calculation of extension theory and the degree of variance among networks is also reflected. Then in the combination of service types obtaining service membership degree of each decision-making factor, the comprehensive weights of each decision-making factor obtained by using Fuzzy Analytic Hierarchy Process (AHP). This mechanism overcame the subjectivity of man-made scoring in AHP. The other is to propose a simple and effective optimal method taking into account of users’and service suppliers’preferences, by which the performance of function value of each candidate network under the preferences of users and service suppliers would be calculated respectively by Simple Additive Weigting(SAW) method, and corresponding weighting coefficient could be dynamically set in accordance with the actual situation of network. Then adding the performance values of each candidate network together will achieve the general performance value of each candidate network, and the one with maximum general performance value is optimal network. Simulation results show that the proposed network selecting mechanism, on the premise of taking users’ and service suppliers’ preferences into consideration, effectively balanced the load of different network and increased the use efficiency of the whole system.In the second place, an optimized idea is proposed taking handoff trigger in the process of vertical handoff of heterogeneous wireless network, target network selecting and cutover execution as an organic whole. Taking SINR as the decision factor of handoff trigger, a combined model prediction method is proposed by connecting gray prediction model and BP neural network, so as to increase the predicting accuracy of SINR and reliability of switch trigger mechanism. Then a cost function is constructed, which could select target network in line with network actual situation, requirement of QoS and users’preference. Before the execution of vertical handoff, the selected target network would be accessed several times in regular period, so as to reduce the unnecessary handoff and system signaling overhead. Simulation results show that the proposed vertical handoff mechanism is capable of increasing the system throughput, reducing the number of times of vertical handoff and lowering system dropping probability.Next, an optimal combined call admission control mechanism is proposed, in which the peak rate that users could get in different areas covered by wireless network and different service types are taken into consideration, and the joint call control mechanism is modeled into a semi-Markov decision-making model. Corresponding status spaces are established based on network capacity, and corresponding action spaces are also established in view of different service types and call types, at the same time, state transition probability of system is deduced on the basis of service model. In addition, in the light of QoS requirement and analysis of network conditions, an admission utility function based on service types is put forward to estimate theadmission utility. On the basis of guaranteeing success rate of service admission and low blocking rate, the maximum value of average admission utility will be obtained through optimal control strategy. Simulation results show that the established admission utility function is able to effectively reflect the influences of changing network conditions to admission control, satisfy the requirement of QoS, and obtain higher average admission utility and larger system throughput and guaranteeing lower blocking rate.Final, this paper proposes prcing and resource allocation scheme based on game theory in multi providers and multi users heterogeneous wireless environment. First discusses the noncooperative game producing effect on provider pricing, Nash equilibrium is considered as the solution of this game. Next, considers the pricing under the cooperative game, comparing the providers revue difference between noncooperative game and cooperative game. Final, a allocation scheme is adopted to attain maximation revue of both providers ang users based on multi-leader Stackelberg games, and designs a user revue function about uer utility, user cost and delay cost. A distributed iterative algorithm is derived to obtain optimal bandwidth allocation and pricing strategies for user and providers respectively.更多还原