基于改进遗传算法的网络资源优化配置方法分析-计算机应用技术专业论文.docxVIP

摘 要 随着因特网的发展，越来越多具有 QoS 要求的业务接入因特网。网络规模的 不断扩大以及业务种类的日趋多样化，使得业务 QoS 需求与资源消耗代价之间的 矛盾日益加剧。如何去协调两者之间的矛盾，网络的资源优化配置工作就显得尤 为重要。带宽与节点缓冲区是网络系统的两类重要资源，两者的合理配置直接影 响着网络系统的性能以及业务的服务质量。由于现有的资源配置方法，大多是针 对单一资源的配置管理，没有考虑多类资源的综合配置对业务 QoS 的影响，容易 造成资源的浪费；再者，单一资源的优化配置算法不能有效的保证业务 QoS 需求 问题，易造成资源代价增加。因此，研究带宽与缓冲区两种资源的合理配置机制 是非常必要的。 本文在充分分析了网络业务特征、网络系统特征的基础上，建立了一个基于 业务时延、丢包率、业务量、带宽以及缓冲区等参数的资源优化配置模型，并提 出了一种求解该模型的方法，该方法通过对传统的遗传算法进行改进，采用基于 向量的编码机制和精英保留的选择策略，解决了传统遗传算法收敛速度慢、易于 陷入局部最优解的问题。最后，通过 OPNET 仿真模拟，验证了本文优化配置方法 的正确性与有效性。 关键词： 网络优化，遗传算法， QoS Abstract With the development of the Internet, more and more business with QoS requirements of Internet access. The continuous expansion of the network size and become more diverse types of business, making business QoS resource consumption demand and the contradiction between the cost of growing. How to co-ordinate between the two conflicts, optimize the allocation of network resources is particularly important work. Bandwidth and node buffer is a system of two important types of network resources, the rational allocation of both a direct impact on the performance of network systems, as well as the quality of the services business. The allocation of resources as a result of the existing methods, most of the resources for a single configuration management, does not take a wide range of resources integrated QoS configured on the impact of business and creating a waste of resources; Furthermore, a single algorithm for optimal distribution of resources is not an effective guarantee QoS needs of business and easy to increase the cost of resources. Thus, two types of bandwidth and buffer mechanism for the rational allocation of resources is essential. In this paper, a full analysis of the characteristics of the network operations, network systems and operational characteristics of QoS and network resources based on the mapping, the establishment of a business based on latency, packet loss rate, volume, bandwidth and buffer parameters such as optimal allocati