We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

nlin.CD

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Nonlinear Sciences > Chaotic Dynamics

Title: Network Reconstruction Problem for an Epidemic Reaction-Diffusion

Abstract: We study the network reconstruction problem for an epidemic reaction-diffusion. These models are an extension of deterministic, compartmental models to a graph setting, where the reactions within the nodes are coupled by a diffusion. We study the influence of the diffusion rate, and the network topology, on the reconstruction and prediction problems, both from a theoretical and experimental standpoint.Results first show that for almost every network, the reconstruction problem is identifiable. Then, we show that the faster the diffusion, the harder the reconstruction, but that increasing the sampling rate may help in this respect.Second, we demonstrate that it is possible to classify symmetrical networks generating the same trajectories, and that the prediction problem can still be solved satisfyingly, even when the network topology makes exact reconstruction difficult.
Subjects: Chaotic Dynamics (nlin.CD)
Cite as: arXiv:2109.11334 [nlin.CD]
  (or arXiv:2109.11334v1 [nlin.CD] for this version)

Submission history

From: Pierre-Yves Masse [view email]
[v1] Wed, 22 Sep 2021 08:25:31 GMT (549kb,D)

Link back to: arXiv, form interface, contact.