论文部分内容阅读
QoS路由问题作为多目标约束优化问题,是一种非确定多项式完全(NP-complete)问题,目前在地面网络中多使用启发式算法求解。卫星网络的拓扑时变特性要求路由算法快速收敛,而高昂的信息交换代价又要求尽量减少星间控制信息交换,这导致绝大多数探测导向型的启发式路由算法应用在星上时性能不高。该文引入一种基于正交多项式神经网络的卫星网络QoS路由算法,将数据包路由过程类比为在经训练过的神经网络中分类的过程;同时正交多项式的使用提高了训练速率,保证了拓扑周期内训练结果的有效性。仿真结果表明:该路由算法在满足用户QoS需求的同时还降低了链路拥塞、丢包率、呼叫阻塞率等指标。
As a multi-objective constrained optimization problem, QoS routing problem is a non-deterministic polynomial NP-complete problem. Currently, many heuristic algorithms are used to solve this problem in terrestrial networks. The time-varying topological characteristics of satellite networks require fast convergence of routing algorithms, and the high information exchange cost requires minimizing the exchange of interstellar control information. This leads to the vast majority of probe-oriented heuristic routing algorithms that perform poorly on star . This paper introduces a satellite network QoS routing algorithm based on orthogonal polynomial neural network, which classifies the data packet routing process as the process of classification in the trained neural network. At the same time, the use of orthogonal polynomials improves the training rate, Effectiveness of training results in topological cycles. Simulation results show that the proposed routing algorithm can meet the QoS requirements of users while reducing the link congestion, packet loss rate, call blocking rate and other indicators.