Adaptive Control Technology Research Based On The Neuron

With the popularity of the Internet, the service provided apparently already cannot satisfy theuser and the business requirement for the quality of different services. Thus improving theutilization ratio of network resources and providing users with higher quality of service is the goalof the dynamic research field. But there still exists a lot of active queue algorithm which is sensitiveto parameter setting and cannot accommodate the dynamic change of network, provides higherrobustness and stability.This paper studies the active queue management algorithm is RED and ARED, Gentle-RED,RIO-C algorithms, such as evaluate the advantages and disadvantages of each algorithm, focus onclassic control algorithm, and join control fuzzy control and neural network to propose an improvedalgorithm. It can protect high-priority data to solve the problem of parameter sensitivity, andmaintain stability. The algorithm uses fuzzy control module, the proportional coefficient of singleneuron online adjustment, and the learning rate of the single neuron has carried on thecorresponding improvement. Using NS2to simulate the performance of improved algorithm itself,at the same time comparing with the traditional PID algorithm, simulation results show that theimproved algorithm has better convergence and stability.