A Framework of Robust Congestion Mitigation with Traffic Estimation, Split Ratio Optimisation and Path Planning

Extensive traffic micro-bursts as well as limited capabilities of network infrastructures result in unpredictable network congestions which have large impacts on the quality of service. In order to effectively and robustly mitigate network congestions, this paper proposes a framework which incorporates future traffic estimation with edge learning, split ratio optimisation via the multi-step linear programming, as well as path planning using a revised column generation algorithm. The combination of traffic prediction and the proposed optimisation framework takes future traffic into consideration and guarantees the robustness of segment routing process. The revised column generation algorithm enables flexible re-routing of traffic distributed to congested links and significantly reduces congestion areas. Extensive experiments on a specific traffic dataset and three benchmark datasets show the feasibility and effectiveness of the proposed algorithms.