Network Congestion Diffusion Model Considering Congestion Distribution Information

Abstract

Network congestion diffusion has become the most stubborn disease and scourge of networks. With the help of widely used congestion distribution information, making full use of network capacity dynamically is a feasible and hopeful way to alleviate network congestion diffusion. Existing studies on network congestion diffusion considering congestion distribution information mainly focus on road networks and describe the distribution information of congestion areas with statistical parameters but not take dynamical congestion distribution into account. However, it is difficult to quantify dynamical congestion distribution as it has multiple influencing factors and complex dynamical coupling relationships, and thus there is still a lack of common network congestion diffusion model considering dynamical congestion distribution information. Inspired by the Langevin diffusion model in signal transduction networks, we propose a novel model for common network congestion diffusion considering the influence of dynamical congestion distribution information based on a set of differential equations. In these equations, we quantify the crosstalk influence of dynamical distribution information by a parameter with reference to the routing optimization method in the ant colony algorithm. And, then firstly the complex dynamical coupling network congestion diffusion under the influence of congestion distribution information is analyzed and simulated in a measurable way. The simulation results prove that there are obvious alleviated effects on network congestion diffusion with proper information influence weights, which is shown to be a bathtub curve relationship. Our model provides a simple mathematical approach to discover the relationship between network congestion diffusion and the influence of dynamical congestion distribution information. Based on this relationship, we can relieve network congestion by dynamically adjusting congestion distribution information influence.