复杂网络信息容量优化策略研究

发布时间：2018-04-27 22:32

本文选题：复杂网络 + 信息流动力学　；参考：《哈尔滨工业大学》2016年博士论文

【摘要】：网络最重要的功能之一是有效传输其所承载的对象,而现代社会处在一个信息爆炸的大数据、大流量时代,比如互联网高速发展,网络数据量急剧增长;随着工业发展和经济增长,交通车辆成倍增加;随着工业用电的增加,电力网所承担的负载不断加重。在这种情况下,网络拥塞现象经常发生,网络信息容量不足与数据流日益增长的矛盾越来越突出,如何提高网络信息容量以降低网络拥塞成为一个急需解决的问题。自从20世纪90年代,小世界和无标度网络模型被提出后,复杂网络获得了长足的发展。复杂网络理论成为理解实际网络结构和功能的有力工具,复杂网络动力学行为的研究有助于理解实际网络涌现出来的各种现象。通过分析复杂网络上的信息流动力学过程,提出有效提高网络信息容量的策略,可以有效抑制网络拥塞的发生。研究发现网络信息容量主要与网络结构密切相关,同时还受路由策略和资源分配因素的制约。因此,为了使网络的负载分配更合理,信息流更均匀,重点研究了提高复杂网络信息容量的优化策略:(1)优化网络拓扑结构方面。提出了一种基于介数的边有向化策略,通过把可能负载重的边限制成单向边的方法使得数据包绕行到负载轻的边上,从而实现了网络信息流的均匀化。在同一种有向化策略下,非随机确定方向的方法比随机确定方向的方法更能提高网络的信息容量。边有向化策略虽然导致了数据包传输的平均最短路径长度增加,却能大幅度提高网络信息容量。为了提升网络的整体性能,提出一种基于局部度和节点间最短路径长度的增边策略,与基于度的增边策略相比,在增加同样数量边的情况下,更能有效提升网络的信息容量并优化其他相关的性能参数。(2)优化路由策略方面。对双层复杂网络的结构特点进行了研究,分析了其信息流动力学过程,提出一种基于逻辑层和物理层节点介数的有效路由策略,该路由策略对逻辑层和物理层的边分别设置权重,并选择权重最小的路径作为数据包的传输路径,使得数据包在传输过程中能避开逻辑层和物理层的hub节点,从而促进了网络信息流分布的均匀化,有效提升了网络信息容量。为了平衡hub节点的传输能力并减少计算复杂度,提出一种基于逻辑层和物理层节点度的机动路由策略,该路由策略使得hub节点也有机会传输数据包,提高了节点的利用率,而且此策略是基于网络局部特征节点度选择的数据包传输路径,计算量较小。该机动路由策略可以有效提高网络信息容量,从而缓解了网络拥塞现象的发生。(3)优化资源分配方面。目前大部分关于双层复杂网络信息容量的研究都是平均分配节点的处理能力,这种资源分配方式制约了网络信息容量的提高。在网络节点总的处理能力一定的情况下,合理分配节点的处理能力可以有效提升整个系统的信息容量。提出一种有效分配物理层节点处理能力的策略以优化双层复杂网络的信息容量,该策略根据物理层和逻辑层的度分布来分配物理层节点处理能力。理论分析和仿真结果表明该资源分配策略优于节点处理能力平均分配策略,可以有效提高双层复杂网络信息容量。由于介数更能精确反映网络中节点的负载情况,所以基于逻辑层和物理层节点的介数提出一种节点处理能力分配策略,该策略比前者更能有效提高网络的信息容量,但是由于代价函数基于介数,所以计算量较前者策略会大一些。
[Abstract]:One of the most important functions of the network is the effective transmission of the objects they carry, and the modern society is in a large data explosion of information. In the era of large traffic, such as the rapid development of the Internet, the amount of network data is increasing rapidly; with the development of industry and economic growth, traffic vehicles are multiplied; with the increase of industrial electricity, the power network bears In this case, network congestion often occurs, the contradiction between the lack of network information capacity and the increasing number of data flow is becoming more and more prominent. How to improve the network information capacity to reduce network congestion has become an urgent problem. Since 1990s, the small world and scale-free network model has been proposed, complex. Complex network has made great progress. Complex network theory has become a powerful tool to understand the structure and function of the actual network. The study of complex network dynamic behavior helps to understand the emergence of various phenomena in the actual network. By analyzing the process of information flow mechanics on complex networks, a strategy to effectively improve the network information capacity is proposed. It can effectively suppress network congestion. It is found that the network information capacity is closely related to the network structure, and is also restricted by the routing strategy and resource allocation factors. Therefore, in order to make the network load distribution more reasonable and the information flow is more uniform, the optimization strategy for improving the information capacity of complex network is studied emphatically: (1) the optimization network In the aspect of network topology, an edge oriented strategy based on the number of medials is proposed. By limiting the possible negative load edge to a one-way side, the packet is detorbed to the light side of the load, thus realizing the homogenization of the network information flow. Under the same oriented strategy, the non random direction determination method is compared to the random direction. The method can improve the information capacity of the network. Although the edge oriented strategy leads to the increase of the average shortest path length of the packet transmission, it can greatly improve the network information capacity. In order to improve the overall performance of the network, an edge increasing strategy based on the local degree and the shortest path length between nodes is proposed. In the case of increasing the same number of edges, it can effectively improve the information capacity of the network and optimize the other related performance parameters. (2) optimizing the routing strategy. The structure characteristics of the double layer complex network are studied, the information flow mechanics process is analyzed, and an effective path based on the logic layer and the physical layer node number is proposed. By strategy, the routing strategy sets the weight of the logic layer and the edge of the physical layer respectively, and selects the path of the least weight as the transmission path of the packet, so that the packet can avoid the hub nodes of the logic layer and the physical layer during the transmission process, thus promoting the uniformity of the network information flow distribution and effectively improving the network information capacity. In order to balance the transmission ability of hub nodes and reduce the computational complexity, a mobile routing strategy based on logical layer and physical layer node degree is proposed. The routing strategy makes the hub node also transmit data packets and improve the utilization rate of nodes, and this strategy is based on the packet transmission path of the network Bureau feature node degree selection. The maneuver routing strategy can effectively improve the network information capacity and alleviate the occurrence of network congestion. (3) optimize the allocation of resources. At present, most of the research on the information capacity of the double complex network is the processing ability of the average allocation nodes. This resource allocation method restricts the network information capacity. In the case of the total processing ability of the network nodes, the reasonable allocation of the node's processing ability can effectively improve the information capacity of the whole system. A strategy to effectively allocate the processing capability of physical layer nodes is proposed to optimize the information capacity of the double complex network. The strategy is allocated according to the degree distribution of the physical layer and the logic layer. The theoretical analysis and simulation results show that the resource allocation strategy is superior to the node processing capability average allocation strategy, which can effectively improve the information capacity of the double layer complex network. Because the mediator can more accurately reflect the load situation of the nodes in the network, a node based on the number of logical layer and physical layer nodes is proposed. The strategy of processing power allocation is more efficient than the former, but because the cost function is based on the number of medials, the amount of computation is larger than that of the former.

【学位授予单位】：哈尔滨工业大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：O157.5

【相似文献】