迭代学习在网络控制中的应用*

针对网络拥塞控制中网络拥塞本身无法建立精确的数学模型的问题，基于迭代学习控制具有结构简单及对系统精确模型不依赖等优点，首次提出了用迭代学习控制算法来解决网络拥塞，其主要目的是提高网络资源的利用率并提供给信源公平的资源分配份额。在提出算法前，首先通过分析网络模型建立了网络拥塞被控系统；然后提出了针对该被控系统的开闭环PID型迭代学习控制算法并证明了其收敛性；最后运用此算法建立了网络拥塞控制模型。通过实验和仿真表明，该算法对解决网络拥塞问题有很好的效果。

Applying iterative learning to network control

With the goal of improving the network resource utilization rate and supplying an impartial resource ditribution quota for source,this paper proposed the solution of iterative learning control algorithm to the network congestion problem by ma-king use of the merits of iterative learning control such as simple structure and independence of accurate system model.Prior to describing the algorithm,it constructed a controlled system for the network congestion by analyzing network model,and then gave the controlled system an open-closed loop PID type iterative learning control algorithm and proved its convergence,finally built a network congestion control model using the algorithm.Experiment and simulation prove that this algorithm takes a good effect in solving the network congestion problem.