Research Article

Dynamic Request Routing for Online Video-on-Demand Service: A Markov Decision Process Approach

Figure 3

An example to find departure probability with a small . , . The upper bound of sojourn time is . We obtain that in this scenario.
920829.fig.003