Method for realizing fuzzy neuron active queue management method in IPCOP

Abstract

The invention discloses a method for realizing a fuzzy neuron active queue management method in IPCOP. The method comprises the following steps of: 1, defining and initializing related parameters; 2, waiting for a new data packet; 3, dynamically adjusting the neuron weighting coefficient by use of a derived Hebb learning algorithm, including the steps of sampling the instant queue length q(k) at current moment and calculating the values of the input variables x1(k), x2(k) and x3(k) of the neuron method; 4, dynamically adjusting the neuron grain K by use of a fuzzy control method, including a step of fuzzifying the input variables x1(k) and x2(k) by taking the x1(k) and x2(k) as input and the variation (delta)K of the neuron gain K as output; and 5, calculating the discard probability p(K) and discarding the data packet according to the discard probability p(k), including the steps of calculating the discard probability p(k) and finally discarding the packet according to the discard probability p(k).

Description

technical field [0001] The invention relates to a method in which a fuzzy neuron active queue management method in an actual communication network experiment platform is realized in an IPCOP software router. Background technique [0002] In recent years, computer network has been widely used as a new technology, and many researchers have done a lot of research to improve its performance. However, with the continuous development of science and technology, the scale of the network has been growing. The excessive demand for limited network resources has caused the congestion of the computer network, resulting in a large loss of network data packets, low utilization of network resources, and some other performance deterioration. . Therefore, network congestion, as a relatively serious problem, has attracted the attention of researchers. [0003] It is far from enough to rely solely on source-end control strategies and mechanisms for control. IETF (Internet Engineering Task For...

AI Technical Summary

Problems solved by technology

In order to improve the overall performance of the system, a differential link is added on the basis of PI. For example, PD and PID appear as an AQM algorithm to adjust the queue length. However, due to the delay and randomness of the computer network itself, these The handling effect of the controller is not very ideal

Moreover, it is problematic and impractical to treat the network as a linear time-invariant system, because the actual network is complex and changeable

Therefore, these methods have limited effectiveness, and these algorithms are implemented in NS2 simulation software, and have not been applied in the actual network environment

Images