Benders Decomposition-based video bandwidth allocation in mobile media cloud network

Abstract

Due to the fast growth of wireless multimedia applications, mobile media cloud network is getting more and more popular. In the architecture of mobile media cloud network, the wireless access points are placed on edge of the cloud to provide media services for the mobile users. The video bandwidth allocation managed by a centralized media cloud directly affect the user’s experiences. In this paper, the problem of the video bandwidth allocation in the mobile media cloud access network is explored. Firstly, this paper formulates the problems in bandwidth allocation in the form of quadratic programming in order to maximize the system revenue on the basis of video bitrates capacity between the user and the Mobile Access Edge Point (MAEP). The optimization model could more vividly explicate the trade-off between the expected bitrates capacity and the allocation fairness of User Equipment (UE). Then this paper subdivides the problem into major and minor ones and proposes an algorithm based on Benders’ Decomposition to deal with it. The optimality of the solution is proved by both theoretical and experimental investigations. The error tolerance is analyzed as the algorithm disavows the trade-off between the convergence time and the system performance. The experiments show that the average computing time and confidence interval of the proposed algorithm are lower than Simplex algorithm by 68% and 94% and Barrier algorithm by 46% and 75% respectively at most. Finally, some conclusions are derived from evaluations on the system performance against various network topologies and different values for parameters of the proposed algorithms.