Research On Implementation Of PID AQM Algorithm In The Real Network And Data_driven Mode

With the increasing development of the Internet and the new network applications, the network flow is rapidly increasing, and network congestion became the bottleneck of the development and application in the Internet. The Internet congestion control mechanism consists of AQM in the routers and TCP protocol end‐to‐end congestion control algorithms. The aim of the AQM is lower queue delay and higher data throughput by dropping randomly IP packets in the intermediate nodes.Currently, researchers have proposed a number of AQM algorithms, and most of all research the algorithms based on RED, later the control theory is introduced into the AQM. But there are few studies on the implementation mode of AQM algorithm. Although the simulation software, such as NS2, OPENNET are very powerful. But after all, they are not a real network environment. For this, this paper presents the implementation of AQM algorithm in the real network, which could exams the performance of AQM algorithm is better than the simulation network platform.With the development of the industry and modern science and technology, the demand for the performance of the control system is higher. The researchers begin to seek the new theories and methods which are low requirement on the model, convenience of online calculation, good integrated control effect. So the data‐driven control method is proposed and developed. The control engineers value the data‐driven control method, because only the system I/O data is needed to complete the controller design. This paper introduces data‐driven control method to the AQM algorithm, which lays a foundation for the more complex predictive control theory being applied to the AQM.The innovative works of this paper are as follows:1. This paper describes the status and the design idea of the real network and data‐driven, and the mode of AQM implementation. So far, most of the AQM algorithms are based on the model‐driven mode, and the people make a research on the simulation platform. This paper proposes the new implementation mode of AQM algorithm: the implementation of the AQM algorithm in the real network and data‐driven mode.2. This paper makes a careful research on the performance of PID algorithm in the different environment. The results reveal that the performance of PID is better with the high load or the changing of the load. And PID responds faster than the RED controller while PID is superior in robustly regulating the steady‐state value of the queue level. 3. This paper describes the real network environment. The real network which satisfies the normal communication of network and also be set the parameters is designed.4. This paper describes the method of the implementation of PID AQM algorithm embedded in the software router IPCop in the real network in details. And the embedding method is proved to be correct through the experiment. It means the AQM algorithm is applied to the real network environment.5. In this paper, the data‐driven mode is preliminarily applied to the AQM, the PID based on the data‐driven AQM is proposed. The results of the PID in the data‐driven mode are similar to the PID in the module mode in NS2 simulation, so the realization that data‐driven is applied to AQM algorithm is successful.