Design a robust controller for congestion avoidance in TCP/AQM system. (February 2023)

• Record Type: Journal Article
• Authors: Abood, Layla H.; Oleiwi, Bashra Kadhim; Humaidi, Amjad J.; Al-Qassar, Arif A.; Al-Obaidi, Abdulkareem Sh. Mahdi
• Is Part Of: Advances in engineering software. Volume 176(2023)
• Publication Date: 2023-02
• Subjects: Computer network -- Congestion control -- PI controller -- TCP/AQM -- WOA; Computer-aided engineering -- Periodicals; Engineering -- Computer programs -- Periodicals; Engineering -- Software -- Periodicals; Periodicals; 620.0028553
• DOI: 10.1016/j.advengsoft.2022.103395
• Languages: English
• ISSNs: 0965-9978

Abstract

Highlights: Circulating of data causes network congestion, which results in packets time delay, loss and butter overflow. As a solution, an adaptive controller is suggested to achieve robust AQM system using Whale Optimization Algorithm. The purpose is to enhance the system performance and test its efficiency and superiority. The results reflect an efficient and robust response. The obtained results indicate the robust controller behavior and its adaptability to achieve stable system response. Abstract: The development of the life, increasing the internet users and circulated data on the computer network caused the network congestion. Accordingly, some essential issues arose in terms of packets time delay, packets loss, and buffer overflow in the intermediate routers. In order to tackle those issues, Transmission Control Protocol and Active Queue Management (TCP/AQM) has been introduced. AQM is the main approach used to control congestion and overcome those problems to improve network performance. In this study, an adaptive controller is suggested to achieve robust AQM system based on a smart Whale Optimization Algorithm (WOA) to enhance the system performance, then to test its efficiency and superiority with other controllers a comparison with two classical PI controller is done which indicates that the suggested controller is more desirable performance than other controllers. The simulation of TCP/AQM are carried out using MATLAB (R2020a), simulation results reflect an efficient and robust response in facing different matters may happens during simulation like change model parameters (±20%) or facing disturbance signal, in all types of robustness analysis system need small duration time to return to its desired response, 15 s when changing the no. of sessions and the no. of packets and need 3 s when adding the disturbance signal. The results obtained indicate the robust controller behavior and its adaptability to achieve stable system response.