A stochastic flow network model with almost-fractional routing

Abstract:

A discrete-time fractional-routing model is defined for the flow of quantal items in a closed network. In contrast to models with independent probabilistic routing of items, the fractional routing model almost exactly dictates the proportions of items at each network node that are transmitted to neighbors. Although the model has a nonlinear structure, the expectations of nodal item counts are shown to be governed by a linear dynamical system, and in fact to be identical to the expected item counts for a probabilistic routing model defined on the same network graph. Further, the fractional routing model is shown to exhibit lower variability as compared to a probabilistic routing model. Finally, statistical analysis of items' locations in the network is briefly discussed, and an example is developed which illustrates the dynamics of the model.

Año de publicación:

2017

Keywords:

Fuente:

scopus