基于粒子群优化的网络拥塞控制新算法

PI控制器常用于主动队列管理中,但参数整定上的试凑法具有盲目性,算法的瞬态性能也不够理想.本文推导了基于流体流理论的网络简化模型,基于该模型将集群智能中的改进粒子群优化算法(PSO)应用于PID控制器参数优化,定义了一个综合调节时间、上升时间、超调量、系统静态误差、正弦跟踪误差等动静态性能指标函数,在给定的参数空间进行组合优化搜索,迅速求得获取使性能指标优化函数极小化的一组PID控制器参数,将PID控制器应用于网络主动队列管理系统中.仿真结果表明,在大时滞和突发业务流的冲击两种情况下,该方法设计的控制器的动静态性能优于RED、PI算法,超调量均小于5%,调节时间分别小于5秒、4秒,稳态误差分别小于两个数据包和3个数据包.

A New Network Cogestion Control Algorithm Based on Particle Swarm Optimization

PI controller is often used to control active queue management(AQM),but its trial method of tuning controller is aimless,and the dynamic performance of algorithm is not enough satisfying.Simplified network model based on fluid flow theory is derived in this paper,and based on this model,an improved algorithm,i.e.particle swarm optimization(PSO) algorithm is applied to optimization of PID controller parameters.In the following,a new performance function including the system adjusting time,rise time,overshoot,steady state error and sinusoidal position tracking error is defined.It is fast to calculate a group of PID controller parameters that minimize the evaluation function by searching in the given controller parameter area,and then the PID controller is applied to AQM system.The simulation experimental results show that under the two conditions of large time delay and sudden business flow,the overshoot is both less than 5%,the adjusting time is less than 5 seconds and 4 seconds separately,and the steady error is less than 2 packets and 3 packets separately,so the dynamic state and steady state performances of the proposed algorithm are obviously superior to those of the existing RED and PI algorithms under the two conditions.