A cocktail of chemical reaction networks and mathematical epidemiology tools for positive ODE stability problems

Florin Avram, Rim Adenane, Andrei-Dan Halanay · 2026

Read A cocktail of chemical reaction networks and mathematical epidemiology tools for positive ODE stability problems by Florin Avram, Rim Adenane, Andrei-Dan Halanay (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-06778.txt.

Introduction

Avram and colleagues establish two key facts about Jacobians for positive ODEs on edge faces. Each face is defined by a siphon of a chemical reaction network. On such a face, the matrix takes block lower-triangular form, with a transversal block for species that fade on the face and an along-face block. That block is a Metzler matrix. Those facts prove and widen the Next Generation Matrix theorem and make its assumptions clear. They show a boundary fixed point cannot turn unstable through a Hopf bifurcation crosswise.

Catalog record

CREATORFlorin Avram, Rim Adenane, Andrei-Dan Halanay
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE59365
LICENSEhttp://creativecommons.org/licenses/by-sa/4.0/
RIGHTSCC BY SA
SOURCEarXiv

Source: https://arxiv.org/abs/2603.06778

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A cocktail of chemical reaction networks and mathematical epidemiology tools for positive ODE stability problems Florin Avram, Rim Adenane, Andrei-Dan Halanay Abstract We continue recent attempts to put together concepts and results of chemical reaction networks theory (CRNT) and mathematical epidemiology (ME), for solving problems of stability of positive ODEs. We provide first an elegant chemical reaction networks-flavored generalization of the most cited result in ME, the next generation matrix (NGM) theorem. We review next the “symbolic-numeric" approach of Vassena and Stadler, which tackles bifurcation problems by viewing the characteristic polynomial of the Jacobian at fixed points as a formal polynomial in the “symbolic reactivities", and identifies its coefficients as “Child Selection minors of the stoichiometric matrix". We also review two applications of this approach using the Mathematica package Epid-CRN at https://github.com/florinav/EpidCRNmodels , which implements tools from both chemical reaction networks and mathematical epidemiology. MSC Classification] 34C10, 34D23, 34C25, 92C45,92C60 keywords biochemical interaction networks; essentially nonnegative/positive systems; mathematical epidemiology; disease free equilibrium; regular splitting; chemical reaction networks; stoichiometric matrix; siphons/semi-locking sets; autocatalytic cores; symbolic-numeric Jacobian; symbolic-numeric Child Selection; rich kinetics; Blokhuis-Stadler-Vassena oscillation recipes; SIR model; nonlinear force of infection; bifurcation analysis; periodic solutions pacs [ 1 Introduction This essay originated in a paper on the importance of reproducible scientific computation Adenane et al. 2023 , where we provided a correction of a famous paper (294 citations, and growing) concerning Hopf and Bogdanov-Takens bifurcations due to L. Zhou and M. Fan, “Dynamics of an SIR epidemic model with limited medical resources revisited", in which we discovered and reported a significant numerical omission in Figure ( Zhou and Fan 2012 , Fig. 6) . In the time since, our computational efforts lead us to discover the parallel chemical reaction networks (CRN) literature, with its numerous computing packages, to develop a Mathematica package Epid-CRN at https://github.com/florinav/EpidCRNmodels , which implements tools from both chemical reaction networks and mathematical epidemiology, to throw new light on the Zhou-Fan paper in Vassena et al. 2024 , using the Vassena-Stadler “Child Selections of the symbolic Jacobian" approach, and finally to establish a CRN-flavored generalization of the most cited result in ME, the next generation matrix (NGM) theorem. Contents We start the paper by a brief glimpse on the theory of positive systems in Section 2 . We proceed in Section 3 with our generalization of the NGM theorem. Section 4 reviews the main ideas of the Vassena-Stadler bifurcation analysis of symbolic Jacobians via Cauchy–Binet Child Selections, and illustrates it on the SIRWS ME model. In section 4.5 , we review and comment the recent idea of Boros and Rost 2026 of proving symbolically the presence of Hopf bifurcations for complicated ME models by examining reduced subnetworks and by applying “inheritance rules" Banaji 2018 ; Banaji et al. 2025 . Section 5 reviews results of Vassena et al. 2024 on a general “Capasso-Ruan-Wang" “semi-parametric -semi functional" SIRS-type epidemic model with nonlinear forces of infection and treatment ( 8 ) (which generalizes Zhou-Fan). 2 A bird’s eye view of positive ODEs Population dynamics, ecology, mathematical epidemiology (ME), virology, rumour spreading, social networks, chemical reaction networks theory (CRNT), are a few of the biological interaction networks (BIN) subfields, which study all positive dynamical systems, and have similar preoccupations: the existence and multiplicity of equilibria, their local and global stability, the occurrence of bifurcations, persistence, permanence, extinction, etc. Definition 1 (Positive / non-negative ODE) . An ODE is called positive Rantzer 2015 or non-negative Haddad and Chellaboina 2011 if the non-negative orthant \mathbb{R}^{n}_{\geq 0}:=\{x\in\mathbb{R}^{n}:x_{i}\geq 0,\;i=1,\dots,n\} is forward invariant under the flow. A polynomial or fractional positive ODE may always be represented as: \dot{x}=\Gamma v(x,\kappa),\qquad x\in\mathbb{R}^{n}_{>0},\ \kappa\in\mathbb{R}^{m}_{>0}, (1) where \Gamma\in\mathbb{R}^{n,m} is called the stoichiometric matrix and v(\cdot,\kappa) is called the vector of rates. The rank of the representation is \operatorname{rank}(\Gamma) . In CRN, \Gamma is further decomposed as \Gamma_{kj}=\tilde{s}_{k}^{j}-s_{k}^{j}, (2) where s_{k}^{j},\tilde{s}_{k}^{j}\geq 0 are called reactant and product coefficients, respectively. Thus, \Gamma=\beta-\alpha, where both the reactant and product matrices contain only non-negative elements. Our new results, package and open problems come from the following circle of ideas: 1. For positive ODEs, the boundary faces defined by the zero coordinates of “reasonable" boundary fixed points in the sense of Angeli et al. 2007 are invariant (and may be efficiently computed, by algorithms developed in the theory of Petri nets, under the name of siphons. 2. The minimal siphons correspond often to epidemiologic strains . 3. The “infected variables" at the disease free equilibrium (DFE) are the union of all minimal siphons, and the associated face is the intersection of all invariant boundary faces. 4. The result of Diekmann et al. 1990 ; Van den Driessche and Watmough 2002 ; Van den Driessche and Watmough 2008 may be decomposed in two: (a) A lower triangular structure of the Jacobian on any siphon face, which implies the reduction of the stability problem to that of stability of its diagonal blocks. (b) The existence of regular splittings Jac=F-V Varga 1962 , which allow defining next generating matrices, and characterize the stability at any fixed point on a siphon by a (non-unique) inequality R_{0}=\rho(FV^{-1})<1 . 5. The positive eigenvalues of the NGM on siphons, expressed as functions of the resident (non-siphon) variables, called reproduction functions in Avram et al. 2024 ; Avram et al. 2024 ; Avram et al. 2025 , may also characterize existence and invasion domains of boundary fixed points; this is an observation; for which a precise statement has not yet emerged, and it can be made also for the semi-parametric -semi functional model from the last two sections – see ( 7 ). 6. “Detailed Routh-Hurwitz" bifurcation analysis via Cauchy–Binet Child Selections – see Section 4 , which reviews some ideas from Vassena 2023 ; Vassena and Stadler 2024 ; Blokhuis et al. 2025 ; Golnik et al. 2025 . 3 Generalization of the NGM theorem: Boundary-face invariance forces J_{xy}=0 for polynomial mass-action fields Theorem 1 (NGM theorem for forward invariant faces/siphons) . (generalizing Diekmann et al. 1990 ; Van den Driessche and Watmough 2002 ; Van den Driessche and Watmough 2008 , Johnston and Avram 2025 ) Let (x,y)\in\mathbb{R}^{m}_{\geq 0}\times\mathbb{R}^{n}_{\geq 0} and let \dot{x}=f_{x}(x,y),\qquad\dot{y}=f_{y}(x,y) be a polynomial positive (equivalently, mass-action representable) vector field. Assume that the face F_{x}:=\{x=0\} is invariant , i.e. f_{x}(0,y)\equiv 0\quad\text{for all }y\in\mathbb{R}^{n}_{\geq 0} (3) (equivalently, x is a siphon). Then: 1. The mixed Jacobian block J_{xy}:=D_{y}f_{x}(0,y) vanishes identically on the face F_{x} : D_{y}f_{x}(0,y)=0\quad\text{for all }y\in\mathbb{R}^{n}_{\geq 0}, and at any point (0,y^{\ast}) on the face, the Jacobian has block lower-triangular form J(0,y^{\ast})=\begin{pmatrix}J_{x}&0\\ *&J_{y}\end{pmatrix}. 2. A fixed point on the face F_{x} is stable iff the Jacobians J_{x},J_{y} at that point are both stable. 3. If furthermore J_{x}=F-V , where (F,V) is a regular splitting (i.e. F,V^{-1} are non-negative), then a fixed point on the face F_{x} is stable if \rho(FV^{-1})<1 , and unstable if \rho(FV^{-1})>1 Varga 1962 . Proof. Fix i\in\{1,\dots,m\} and write the i th component of f_{x} as a polynomial (f_{x})_{i}(x,y)=\sum_{\alpha\in\mathbb{N}^{m},\;\beta\in\mathbb{N}^{n}}c_{\alpha,\beta}\,x^{\alpha}y^{\beta},\qquad x^{\alpha}:=\prod_{p=1}^{m}x_{p}^{\alpha_{p}},\quad y^{\beta}:=\prod_{q=1}^{n}y_{q}^{\beta_{q}}, with real coefficients c_{\alpha,\beta} and only finitely many nonzero terms. Evaluating at x=0 yields (f_{x})_{i}(0,y)=\sum_{\beta\in\mathbb{N}^{n}}c_{0,\beta}\,y^{\beta}, because every monomial with \alpha\neq 0 vanishes when x=0 . By the invariance hypothesis ( 3 ), (f_{x})_{i}(0,y)\equiv 0 as a polynomial in y on the set \mathbb{R}^{n}_{\geq 0}

Read the complete text