一种基于RBF网络的参数自调整REM算法

针对传统的随机指数标记（Random Exponential Marking，REM）算法存在参数难以调整的缺陷，提出了一种改进型主动队列管理算法—基于RBF网络的参数自调整REM算法。利用RBF神经网络动态地对算法参数进行调整，使其能够适应不断变化的网络环境。该算法将REM算法在一定条件下近似为比例积分控制，使REM算法参数调节近似等效于比例积分系数的调节，简化了参数调节，提高了算法的实时性。仿真实验表明：该算法优于传统的REM算法，具有良好的鲁棒性及较快的调节速率。

Parameter Self-regulation REM Algorithm Based on RBF Neural Network

Aiming at the shortcoming of random exponential marking (REM) algorithm, whose parameters are difficultly regulated, this paper proposes a new RBF neural network based REM algorithm in which REM parameters are tuned via RBF network technique. By approximating the REM algorithm via PI control under certain conditions, the regulation of REM parameters is equivalent to adjust the coefficients P and I. Thus, the process of regulation may be simplified and the real time control performance may be promoted. The simulation results show that the proposed algorithm can effectively adapt to dynamic network environments, and attain quicker adjusting.