Online optimal queue scheme for enhanced transport over wireless links

In this paper we address the performance corruption in TCP flows over a wireless link by employing an online optimal control scheme. It has been shown that TCP flows exhibit low data rate in wireless scenario due to fading channel, frequent hands off and high bit error. The high bit error rate, which is an intrinsic link characteristic of wireless network, contributes to more packets losses, which cut back the congestion window and decrease the transmission rate. Meanwhile, the unpredictable traffic situations and varying network capacity all lead to dynamics changing of TCP behavior. To overcome these problems, we propose a data-based AQM with an optimality criterion to minimize the fluctuation of the queue length and aggregate rate from link capacity. The adaptive control gain is updated by solving the equivalent Hamiltonian without knowing the systems parameters. The simulation results show that the proposed scheme effectively improves the performance of TCP over wireless link with strong robustness to time-varying scenarios of wireless networks.