Flow control as a stochastic control problem with incomplete information

Boris Miller

Institute for Information Transmission Problems


The nonlinear stochastic control problem related with the Internet flow control is considered. The state of the link is described by hidden Markov process while the loss flow is described by the counting process with the intensity depending on the current transmission rate and unobserved link state. The control is the transmission rate that has to be chosen as non-anticipating process depending on the observation of the loss process. The aim of the control is to maximize some performance criterion characterizing the quantity of successively transmitted infromation and losses. We state the problem with incomplete information where the flow intensity plays a dual role. On one hand it of course has an impact on transmission rate and on the other hand provokes the arrival of the information about the link state via the loss process. It is a new manifestation of the incompelete infromation stochastic control problem which admits a reduction to the stochastic control problem with complete observations and sometimes even to a deterministic control problem. We overview various statements of this problem that could be interesting in the context of modelling and optimization of the IP wired and wireless networks and provide a more detail analysis of one particular setting.

[Boris Miller]
[Institute for Information Transmission Problems]