用于求解多约束QoS路由优化问题的改进伊藤算法  

Improved ITO algorithm for solving multi-constrained QoS routing optimization problems

作　　者：余世明[1] 周凯杰 何德峰[1] Yu Shiming;Zhou Kaijie;He Defeng(College of Information Engineering,Zhejiang University of Technology,Hangzhou 310023)

机构地区：[1]浙江工业大学信息工程学院,杭州310023

出　　处：《高技术通讯》2021年第1期12-20,共9页Chinese High Technology Letters

基　　金：国家自然科学基金(61773345);浙江省自然科学基金(LR17F030004)资助项目。

摘　　要：针对伊藤算法(ITO)在大规模网络中求解多约束服务质量(QoS)路由优化时,存在收敛速度过慢、易陷入局部最优解从而导致算法成功率不高等问题,提出基于多策略协同优化的改进伊藤算法。该算法通过改进漂移与波动过程的结合方式,提出了一种新的协同更新策略,并引入双重认知策略和多精英引导学习策略,设计了一种新的路径权重更新规则。该规则使算法中漂移粒子和波动粒子强度根据个体适应度灵活变化,具有自适应性。仿真结果表明,该算法在保证系统稳定性的基础上,降低了QoS路由的迭代次数与费用,并且在较大规模网络中有理想的表现。The available ITO algorithms have the problems of slow convergence and easily falling into some local optimal solutions during solving a multi-constrained quality of service(QoS) routing optimization problem, which leads to the low success of the algorithms. Aiming at the problems, an improved ITO algorithm is proposed based on mutli-strategy collaborative optimization in this work. By improving the combination mode of drift and wave process, a new collaborative update strategy is presented. Moreover, by introducing a dual cognitive strategy and multi-elite guided learning strategy, a new path weight update rule is designed to raise the strength of the drift particles and make wave particles in the algorithm be flexible and adaptable according to individual fitness. The simulation results show that the proposed algorithm can greatly reduce the iteration times and costs of QoS routing with guarenteed stability of the system, and it has an ideal performance in large-scale networks.

关 键 词：路由优化 服务质量(QoS) 伊藤算法(ITO) 双重认知 协同更新 精英学习 

分 类 号：TP393.0[自动化与计算机技术—计算机应用技术] TP181[自动化与计算机技术—计算机科学与技术]