<u style="text-decoration: none;">Nikki Sonenberg (Antwerp University) - Networks of interacting stochastic fluid models</u>
Stochastic fluid models have been widely used to model the level of a resource that changes over time, where the rate of variation depends on the state of some continuous time Markov process. Latouche and Taylor introduced an approach, using matrix analytic methods and the reduced load approach for loss networks, to analyse networks of fluid models all driven by the same modulating process where the buffers are infinite. We extend the method to networks involving finite buffer models and illustrate the approach by deriving performance measures for a simple network as characteristics such as buffer size are varied. Our results provide insight into the situations where the infinite buffer model is a reasonable approximation to the finite buffer model.