Abstract: In order to mitigate the dropping packet with network congestion problem, a novel passive queue management algorithm （Drop Front n based on Wavelet and Shuffled frog leaping, DFWS） is proposed by drop front. In this algorithm, the long range dependence of actual traffic is depicted by wavelet transform, and the network queue length is deducted with shuffled frog leaping algorithm. Then, drop packet probability and drop packet strategy are presented by drop front. A simulation with actual data is conducted to research on the relationship between network queue length and dropping number, as well as service rate. The results show that, compared with other algorithm, DFWS has better adaptability.

Key words: passive queue management, drop front, drop packet, wavelet, shuffled frog leaping

摘要： 针对因网络拥塞而产生的丢包问题，基于弃头方式提出了一种新的被动队列管理算法（Drop Front n based on Wavelet and Shuffled frog leaping，DFWS）。该算法首先利用小波变换对实际流量的长相关特性进行刻画，并利用混合蛙跳方法获取网络队长，同时结合弃头方式提出丢包策略和丢包概率。最后，以实际数据进行仿真实验，深入研究了网络队长、丢包数量和服务率之间的关系。结果发现，与其他算法相比，DFWS具有较好的适应性。

关键词: 被动队列管理, 弃头, 丢包, 小波, 混合蛙跳