Authors:
Dandan Li
;
Runtong Zhang
and
Xiaopu Shang
Affiliation:
Beijing Jiaotong University, China
Keyword(s):
Cognitive networks, Ant colony algorithm, Neural network, Wavelet, BP (back propagation) neural network, Network traffic prediction.
Abstract:
With the development of the network technology and the increasing demands on communication, more complex, heterogeneous, and suitable network structures are right on their way to come. Cognitive networks can perceive the external environment; intelligently and automatically change its behavior to adapt the environment. This feature is more suitable to provide security for users with QoS. This paper proposes a hybrid traffic prediction model, which trains BPNN with Ant Colony Algorithm based on the analysis of the present models, in order to improve the cognitive feature in the cognitive networks. The proposed model can avoid the problem of slow convergence speed and an easy trap in local optimum when coming up with a fluctuated network flow. At the beginning, the model rejects the abnormal traffic flow data, and then use wavelet decomposition, in the following steps, the model predicts the network traffic with the hybrid model. Thus, the traffic prediction with high-precision in cog
nitive networks is achieved.
(More)