Computing Difference Abstractions of Linear Equation Systems - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Journal Articles Theoretical Computer Science Year : 2021

Computing Difference Abstractions of Linear Equation Systems

Emilie Allart
Joachim Niehren
Cristian Versari
  • Function : Author
  • PersonId : 943051

Abstract

Abstract interpretation was proposed for predicting changes of reaction networks with partial kinetic information in systems biology. This requires to compute the set of difference abstractions of a system of linear equations under nonlinear constraints. We present the first practical algorithm that can compute the difference abstractions of linear equation systems exactly. We also present a new heuristics based on minimal support consequences for overapproximating the set of difference abstractions. Our algorithms rely on elementary modes, first-order definitions, and finite domain constraint programming. We implemented our algorithms and applied them to change prediction in systems biology. It turns out experimentally that the new heuristics is often exact in practice, while outperforming the exact algorithm.
Fichier principal
Vignette du fichier
1.pdf (683.76 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03156136 , version 1 (02-03-2021)
hal-03156136 , version 2 (07-05-2021)
hal-03156136 , version 3 (17-06-2021)

Identifiers

Cite

Emilie Allart, Joachim Niehren, Cristian Versari. Computing Difference Abstractions of Linear Equation Systems. Theoretical Computer Science, 2021, ⟨10.1016/j.tcs.2021.06.030⟩. ⟨hal-03156136v3⟩
160 View
171 Download

Altmetric

Share

Gmail Facebook X LinkedIn More