pith. sign in

arxiv: 2407.19133 · v1 · submitted 2024-07-27 · 📡 eess.SY · cs.SY

Network-Based Epidemic Control Through Optimal Travel and Quarantine Management

Mahtab Talaei , Apostolos I. Rikos , Alex Olshevsky , Laura F. White , Ioannis Ch. Paschalidis This is my paper

Pith reviewed 2026-05-23 23:20 UTC · model grok-4.3

classification 📡 eess.SY cs.SY
keywords epidemic controlSIR modelmatrix balancingquarantineprimal-dual dynamicsreproduction numbernetwork optimizationtravel reduction
0
0 comments X

The pith

Network epidemic control reduces optimal quarantine to a matrix balancing problem that ties directly to the reproduction number.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The authors develop two methods to control epidemics spreading across networks of locations. In the first, they optimize reductions in travel rates between counties, showing that the strategy converges based on how the network is structured. In the second, they augment the classic SIR model with quarantined states and reformulate the problem of choosing quarantine levels as a matrix balancing task. This reformulation creates a clear connection between the optimization constraints and the epidemic reproduction number, which governs whether the disease grows or dies out. They solve the resulting optimization using augmented primal-dual gradient dynamics, which converge exponentially to the optimal solution, and test the ideas on real data from Massachusetts counties.

Core claim

We show that this problem reduces to the problem of matrix balancing. We establish a link between optimization constraints and the epidemic's reproduction number, highlighting the relationship between network structure and disease dynamics. We demonstrate that applying augmented primal-dual gradient dynamics to the optimal quarantine problem ensures exponential convergence to the KKT point.

What carries the argument

The reformulation of the optimal quarantine problem as a matrix balancing problem, which connects the feasible quarantines to the reproduction number.

If this is right

  • The optimal quarantine can be found by solving a matrix balancing problem instead of a general optimization.
  • The constraints ensure the reproduction number stays below a threshold determined by the network.
  • Augmented primal-dual gradient dynamics achieve exponential convergence to the solution.
  • Both approaches are validated through simulations on real county networks.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar matrix balancing reductions might apply to other networked dynamical systems with control inputs.
  • Testing the travel reduction strategy on networks with different connectivity patterns could reveal which topologies allow fastest convergence.
  • The link to reproduction number suggests that quarantine optimization could be used to design interventions that push the effective reproduction number below one.

Load-bearing premise

The epidemic follows a networked SIR model with added quarantine compartments whose rates can be optimized independently of other dynamics.

What would settle it

Observing that the optimal quarantine levels computed via matrix balancing do not keep the reproduction number below one in a real epidemic simulation on the network.

Figures

Figures reproduced from arXiv: 2407.19133 by Alex Olshevsky, Apostolos I. Rikos, Ioannis Ch. Paschalidis, Laura F. White, Mahtab Talaei.

Figure 1
Figure 1. Figure 1: Optimal values of f(τ) in problem (6) for different budget parameters b after changing the travel rates τ via (7). Initial recovered cases are calculated by multiplying the total recovered ratio of US cumulative cases on April 1, 2020 8878 215215 [14] by each node’s cumulative infected population and adding deaths. Active cases, both asymptomatic and symptomatic, are calculated as the remainder of cumulati… view at source ↗
Figure 3
Figure 3. Figure 3: Number of active (asymptomatic and symptomatic infe [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Number of cumulative cases (infected, quarantined, [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Number of active (asymptomatic and symptomatic infe [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
read the original abstract

Motivated by the swift global transmission of infectious diseases, we present a comprehensive framework for network-based epidemic control. Our aim is to curb epidemics using two different approaches. In the first approach, we introduce an optimization strategy that optimally reduces travel rates. We analyze the convergence of this strategy and show that it hinges on the network structure to minimize infection spread. In the second approach, we expand the classic SIR model by incorporating and optimizing quarantined states to strategically contain the epidemic. We show that this problem reduces to the problem of matrix balancing. We establish a link between optimization constraints and the epidemic's reproduction number, highlighting the relationship between network structure and disease dynamics. We demonstrate that applying augmented primal-dual gradient dynamics to the optimal quarantine problem ensures exponential convergence to the KKT point. We conclude by validating our approaches using simulation studies that leverage public data from counties in the state of Massachusetts.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 3 minor

Summary. The manuscript proposes two network-based strategies for epidemic control. The first optimizes reductions in travel rates between nodes, with convergence analysis showing dependence on network structure to minimize spread. The second extends the SIR model to include quarantined states and formulates an optimization problem over quarantine levels; it claims this reduces to a matrix balancing problem, links the constraints to the epidemic reproduction number, and shows that augmented primal-dual gradient dynamics yield exponential convergence to the KKT point. Both approaches are validated via simulations on Massachusetts county-level public data.

Significance. If the claimed reductions to matrix balancing and the exponential convergence results are supported by complete derivations, the work would supply a structured optimization framework that connects standard tools (matrix balancing, primal-dual dynamics) to network epidemic models and reproduction-number constraints. The explicit use of real county data for validation is a strength that grounds the theoretical claims.

major comments (2)
  1. [Abstract and §3] Abstract and §3: The central claim that the optimal quarantine problem reduces to matrix balancing is asserted without an explicit derivation showing how the extended SIR-quarantine dynamics and decision variables map onto the matrix-balancing formulation (including the precise objective and constraint set). This step is load-bearing for the subsequent link to the reproduction number.
  2. [§4] §4: The statement that augmented primal-dual gradient dynamics ensure exponential convergence to the KKT point is given without the Lyapunov function, the explicit step-size conditions, or the rate bound that would establish the exponential claim; these details are required to support the dynamical-systems result.
minor comments (3)
  1. [Notation] The manuscript would benefit from a consolidated table of notation that defines all network, compartment, and optimization symbols in one place.
  2. [Simulations] Simulation section: parameter values drawn from Massachusetts county data should cite the exact public sources and state any preprocessing steps applied to the contact or mobility matrices.
  3. [Figures] Figure captions for the convergence plots should explicitly label the plotted quantities (e.g., primal residual, dual residual, or objective value) and indicate the time scale used.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments. We address each major point below and will incorporate the requested explicit derivations into the revised manuscript.

read point-by-point responses

  1. Referee: [Abstract and §3] Abstract and §3: The central claim that the optimal quarantine problem reduces to matrix balancing is asserted without an explicit derivation showing how the extended SIR-quarantine dynamics and decision variables map onto the matrix-balancing formulation (including the precise objective and constraint set). This step is load-bearing for the subsequent link to the reproduction number.

    Authors: We agree that the reduction claim would benefit from a fully expanded derivation. In the revised manuscript we will add a dedicated subsection in §3 that explicitly maps the extended SIR-quarantine dynamics and decision variables onto the matrix-balancing formulation, including the precise objective, the full constraint set, and the incorporation of the reproduction-number constraint. revision: yes

  2. Referee: [§4] §4: The statement that augmented primal-dual gradient dynamics ensure exponential convergence to the KKT point is given without the Lyapunov function, the explicit step-size conditions, or the rate bound that would establish the exponential claim; these details are required to support the dynamical-systems result.

    Authors: We concur that the exponential-convergence claim requires supporting analysis. The revised §4 will include the Lyapunov function, the explicit step-size conditions, and the derived exponential rate bound that establish convergence to the KKT point. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation applies standard optimization to epidemic model

full rationale

The paper reduces the quarantine problem to matrix balancing and applies augmented primal-dual gradient dynamics to reach KKT points, with a link to the reproduction number derived directly from the extended SIR network model. These steps use established mathematical tools without reducing predictions to fitted parameters by construction, without self-citation load-bearing the central claims, and without ansatzes smuggled via prior work. Simulations on external Massachusetts county data provide independent validation. The derivation chain remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claims rest on the standard assumptions of compartmental epidemic models and graph-based network representations. No free parameters or invented entities are explicitly introduced in the abstract.

axioms (2)
  • domain assumption Epidemic dynamics follow an extended SIR model with quarantined states.
    Core modeling choice invoked for both approaches.
  • domain assumption The underlying contact structure is a fixed network whose travel rates can be optimized.
    Required for the travel-reduction strategy and matrix-balancing formulation.

pith-pipeline@v0.9.0 · 5701 in / 1293 out tokens · 27133 ms · 2026-05-23T23:20:32.517365+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

37 extracted references · 37 canonical work pages · 1 internal anchor

  1. [1]

    Covid-19 dashboard,

    J. Center, “Covid-19 dashboard,” Johns Hopkins University , 2021

    work page 2021

  2. [2]

    Does viral interference affect spread of i nfluenza?

    A. Linde, M. Rotz´ en- ¨Ostlund, B. Zweygberg-Wirgart, S. Rubinova, and M. Brytting, “Does viral interference affect spread of i nfluenza?” Eurosurveillance, vol. 14, no. 40, p. 19354, 2009

    work page 2009

  3. [3]

    Analysis an d control of epidemics: A survey of spreading processes on complex net works,

    C. Nowzari, V . M. Preciado, and G. J. Pappas, “Analysis an d control of epidemics: A survey of spreading processes on complex net works,” IEEE Control Systems Magazine , vol. 36, no. 1, pp. 26–46, 2016

    work page 2016

  4. [4]

    Modeling, estimation, and analysis of epidemics over networks: An overview,

    P . E. Par´ e, C. L. Beck, and T. Bas ¸ar, “Modeling, estimation, and analysis of epidemics over networks: An overview,” Annual Reviews in Control , vol. 50, pp. 345–360, 2020

    work page 2020

  5. [5]

    Optimal co ntrol for pan- demic influenza: The role of limited antiviral treatment and isolation,

    S. Lee, G. Chowell, and C. Castillo-Ch´ avez, “Optimal co ntrol for pan- demic influenza: The role of limited antiviral treatment and isolation,” Journal of Theoretical Biology , vol. 265, no. 2, pp. 136–150, 2010

    work page 2010

  6. [6]

    Multitask le arning and nonlinear optimal control of the covid-19 outbreak: A ge ometric programming approach,

    M. Hayhoe, F. Barreras, and V . M. Preciado, “Multitask le arning and nonlinear optimal control of the covid-19 outbreak: A ge ometric programming approach,” Annual Reviews in Control , vol. 52, pp. 495– 507, 2021

    work page 2021

  7. [7]

    An optimal control problem ove r infected networks,

    A. Khanafer and T. Basar, “An optimal control problem ove r infected networks,” in Proceedings of the International Conference of Control, Dynamic Systems, and Robotics , vol. 125, 2014, pp. 1–6

    work page 2014

  8. [8]

    Optimal control for heterogeneous no de-based information epidemics over social networks,

    F. Liu and M. Buss, “Optimal control for heterogeneous no de-based information epidemics over social networks,” IEEE Transactions on Control of Network Systems , vol. 7, no. 3, pp. 1115–1126, 2020

    work page 2020

  9. [9]

    A closed-loop fra mework for inference, prediction, and control of SIR epidemics on netw orks,

    A. R. Hota, J. Godbole, and P . E. Par´ e, “A closed-loop fra mework for inference, prediction, and control of SIR epidemics on netw orks,” IEEE Transactions on Network Science and Engineering , vol. 8, no. 3, pp. 2262–2278, 2021

    work page 2021

  10. [10]

    Distributed algorit hm for suppress- ing epidemic spread in networks,

    V . S. Mai, A. Battou, and K. Mills, “Distributed algorit hm for suppress- ing epidemic spread in networks,” IEEE Control Systems Letters , vol. 2, no. 3, pp. 555–560, 2018

    work page 2018

  11. [11]

    Convex optimization of the bas ic reproduc- tion number,

    K. D. Smith and F. Bullo, “Convex optimization of the bas ic reproduc- tion number,” IEEE Transactions on Automatic Control , vol. 68, no. 7, pp. 4398–4404, 2023

    work page 2023

  12. [12]

    Optimal resource allocation for network protecti on against spreading processes,

    V . M. Preciado, M. Zargham, C. Enyioha, A. Jadbabaie, an d G. J. Pappas, “Optimal resource allocation for network protecti on against spreading processes,” IEEE Transactions on Control of Network Systems, vol. 1, no. 1, pp. 99–108, 2014

    work page 2014

  13. [13]

    Decreasing the spectral r adius of a graph by link removals,

    P . V an Mieghem, D. Stevanovi´ c, F. Kuipers, C. Li, R. V an De Bovenkamp, D. Liu, and H. Wang, “Decreasing the spectral r adius of a graph by link removals,” Physical Review E , vol. 84, no. 1, p. 016101, 2011

    work page 2011

  14. [14]

    Optimal lockdown fo r pandemic control,

    Q. Ma, Y .-Y . Liu, and A. Olshevsky, “Optimal lockdown fo r pandemic control,” arXiv preprint arXiv:2010.12923 , 2021

    work page arXiv 2010

  15. [15]

    Minimum effort dece ntralized control design for contracting network systems,

    R. Ofir, F. Bullo, and M. Margaliot, “Minimum effort dece ntralized control design for contracting network systems,” IEEE Control Systems Letters, vol. 6, pp. 2731–2736, 2022

    work page 2022

  16. [16]

    Modelling the COVID-19 epide mic and implementation of population-wide interventions in It aly,

    G. Giordano, F. Blanchini, R. Bruno, P . Colaneri, A. D. F ilippo, A. D. Matteo, and M. Colaneri, “Modelling the COVID-19 epide mic and implementation of population-wide interventions in It aly,” Nature Medicine, vol. 26, no. 6, pp. 855–860, 2020

    work page 2020

  17. [17]

    Controlling epid emic spread: Reducing economic losses with targeted closures,

    J. R. Birge, O. Candogan, and Y . Feng, “Controlling epid emic spread: Reducing economic losses with targeted closures,” Management Science, vol. 68, no. 5, pp. 3175–3195, 2022

    work page 2022

  18. [18]

    Viral dynamics of sars-cov-2 infection and the p redictive value of repeat testing,

    S. M. Kissler, J. R. Fauver, C. Mack, C. Tai, K. Y . Shiue, C . C. Kalinich, S. Jednak, I. M. Ott, C. B. V ogels, J. Wohlgemuth, J. Weisberg er, J. DiFiori, D. J. Anderson, J. Mancell, D. D. Ho, N. D. Grubaug h, and Y . H. Grad, “Viral dynamics of sars-cov-2 infection and the p redictive value of repeat testing,” medRxiv, 2020

    work page 2020

  19. [19]

    Synchrony, waves, and spatial hierarch ies in the spread of influenza,

    C. Viboud, O. N. Bjørnstad, D. L. Smith, L. Simonsen, M. A . Miller, and B. T. Grenfell, “Synchrony, waves, and spatial hierarch ies in the spread of influenza,” Science, vol. 5772, no. 312, p. 447–451, 2006

    work page 2006

  20. [20]

    Mul tiscale dynamic human mobility flow dataset in the US during the COVID -19 epidemic,

    Y . Kang, S. Gao, Y . Liang, M. Li, J. Rao, and J. Kruse, “Mul tiscale dynamic human mobility flow dataset in the US during the COVID -19 epidemic,” Scientific data , vol. 7, no. 1, p. 390, 2020

    work page 2020

  21. [21]

    Concerning nonnegative matr ices and doubly stochastic matrices,

    R. Sinkhorn and P . Knopp, “Concerning nonnegative matr ices and doubly stochastic matrices,” Pacific Journal of Mathematics , vol. 21, no. 2, pp. 343–348, 1967

    work page 1967

  22. [22]

    Di stributed weight balancing over digraphs,

    A. I. Rikos, T. Charalambous, and C. N. Hadjicostis, “Di stributed weight balancing over digraphs,” IEEE Transactions on Control of Network Systems, vol. 1, no. 2, pp. 190–201, June 2014

    work page 2014

  23. [23]

    A review of matrix scaling and Sinkhorn's normal form for matrices and positive maps

    M. Idel, “A review of matrix scaling and sinkhorn’s norm al form for matrices and positive maps,” arXiv preprint arXiv:1609.06349 , 2016

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  24. [24]

    Matrix s caling and balancing via box constrained newton’s method and inter ior point methods,

    M. B. Cohen, A. Madry, D. Tsipras, and A. Vladu, “Matrix s caling and balancing via box constrained newton’s method and inter ior point methods,” in 58th Annual Symposium on F oundations of Computer Science, 2017, pp. 902–913

    work page 2017

  25. [25]

    D. G. Luenberger, Linear and nonlinear programming . Klewer Academic Publishers, 2021

    work page 2021

  26. [26]

    Smooth convex appro ximation to the maximum eigenvalue function,

    X. Chen, H. Qi, L. Qi, and K.-L. Teo, “Smooth convex appro ximation to the maximum eigenvalue function,” Journal of Global Optimization , vol. 30, no. 2, pp. 253–270, 2004

    work page 2004

  27. [27]

    On differentiating eigenvalues and eige nvectors,

    J. R. Magnus, “On differentiating eigenvalues and eige nvectors,” Econo- metric Theory , vol. 1, no. 2, p. 179–191, 1985

    work page 1985

  28. [28]

    Quadratic programming,

    J. Nocedal and S. J. Wright, “Quadratic programming,” Numerical optimization, vol. 35, pp. 448–492, 2006

    work page 2006

  29. [29]

    Continuity and location of zeros of linear co mbinations of polynomials,

    M. Zedek, “Continuity and location of zeros of linear co mbinations of polynomials,” Proceedings of the American Mathematical Society , vol. 16, no. 1, pp. 78–84, 1965

    work page 1965

  30. [30]

    Nesterov, Introductory lectures on convex optimization: A basic course

    Y . Nesterov, Introductory lectures on convex optimization: A basic course. Springer Science & Business Media, 2003, vol. 87

    work page 2003

  31. [31]

    Convexity of the dominant eigenvalue of an essentially nonnegative matrix,

    J. E. Cohen, “Convexity of the dominant eigenvalue of an essentially nonnegative matrix,” Proceedings of the American Mathematical Society, vol. 81, no. 4, pp. 657–658, 1981

    work page 1981

  32. [32]

    Semi-global exponential stab ility of augmented primal–dual gradient dynamics for constrained c onvex opti- mization,

    Y . Tang, G. Qu, and N. Li, “Semi-global exponential stab ility of augmented primal–dual gradient dynamics for constrained c onvex opti- mization,” Systems & Control Letters , vol. 144, p. 104754, 2020

    work page 2020

  33. [33]

    2020 Census Data by Count y,

    Massachusetts Legislature, “2020 Census Data by Count y,” https://malegislature.gov/Redistricting/MassachusettsCensusData/County

    work page 2020

  34. [34]

    Gross Domes- tic Product by County, Metro, and Other Areas,

    U.S. Bureau of Economic Analysis, “Gross Domes- tic Product by County, Metro, and Other Areas,” https://www.bea.gov/data/gdp/gdp-county-metro-and-other-areas

    work page

  35. [35]

    Us counties covid 19 dataset,

    e. a. Mitch Smith and, “Us counties covid 19 dataset,” 20 24. [Online]. Available: https://www.kaggle.com/dsv/8581321

    work page arXiv

  36. [36]

    Estimating the fr action of unre- ported infections in epidemics with a known epicenter: An ap plication to covid-19,

    A. Hortac ¸su, J. Liu, and T. Schwieg, “Estimating the fr action of unre- ported infections in epidemics with a known epicenter: An ap plication to covid-19,” Journal of Econometrics , vol. 220, no. 1, pp. 106–129, 2021

    work page 2021

  37. [37]

    Bullo, Lectures on Network Systems , 1.6 ed

    F. Bullo, Lectures on Network Systems , 1.6 ed. Kindle Direct Publishing, 2022. [Online]. Available: https://fbullo.github.io/lns Mahtab Talaei is currently pursuing her Ph.D. in the Division of Systems Engineering at Boston Uni- versity in Boston, MA. She obtained her B.Sc. and M.Sc. degrees in Electrical Engineering from Isfahan University of Technology...

    work page 2022