pith. sign in

arxiv: 2605.20875 · v1 · pith:ARA6IGBGnew · submitted 2026-05-20 · 🧮 math.OC

Spare Strategy for Large-Scale Satellite Constellations Under Dual Resupply Channels Using Markov Chain

Seungyeop Han , Shoji Yoshikawa , Takumi Noro , Takumi Suda , Koki Ho This is my paper

Pith reviewed 2026-05-21 03:48 UTC · model grok-4.3

classification 🧮 math.OC
keywords satellite constellationspare managementMarkov chainhybrid resupplycost optimizationresilience constraintsorbital dynamicsreorder policy
0
0 comments X

The pith

Markov-chain modeling of dual resupply channels yields accurate cost and resilience metrics for satellite spare strategies without aggregation assumptions.

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

This paper develops a Markov-chain method to analyze spare management for large satellite constellations that combine indirect staging in parking orbits with direct deliveries to operating planes. Failures and replenishments are modeled as coupled chains, one using periodic-review reorder rules for the indirect path and a standard reorder policy for the direct path. Solving the chains via fixed-point iteration produces a periodic steady state over the orbital cycle, from which long-run cost and resilience measures are extracted directly. The approach stays valid across wider conditions than earlier aggregated models because it keeps the stochastic multi-echelon dynamics tied to orbital mechanics. A case study confirms the predictions match Monte Carlo runs and shows when the hybrid policy beats using only one channel.

Core claim

The paper claims that satellite failure and replenishment processes modeled as coupled Markov chains—one following a periodic-review reorder-point/order-quantity policy for the indirect channel and a standard reorder-point/order-quantity policy for the direct channel—yield a periodic steady state over the right ascension of the ascending node cycle via fixed-point iteration. Stationary distributions from these chains supply rigorous cost and resilience metrics. An approximate analysis preserves delay statistics while shrinking model size. These metrics support a cost-minimization problem with resilience constraints solved by genetic algorithm; the resulting framework is channel-neutral and,

What carries the argument

Coupled indirect and direct Markov chains iterated to periodic steady state over the right ascension of the ascending node cycle, generating stationary distributions for cost and resilience metrics.

If this is right

  • Stationary distributions supply rigorous long-run cost and resilience metrics for any hybrid policy.
  • Genetic-algorithm optimization finds the lowest-cost design that still meets chosen resilience levels.
  • The approximate analysis keeps delay statistics accurate while cutting model size for quicker evaluation.
  • Because the framework is channel-neutral, the optimizer autonomously assigns roles to indirect and direct paths.
  • The metrics identify the operating regimes in which hybrid resupply outperforms pure direct or pure indirect strategies.

Where Pith is reading between the lines

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

  • The same periodic-steady-state construction could be reused for other repeating orbital effects such as eclipse seasons or radiation exposure cycles.
  • Precomputed distributions could serve as baselines for real-time spare allocation rules that react to observed launch delays.
  • The modeling style might transfer to related multi-stage space logistics questions such as propellant depot management or crew transport scheduling.
  • Direct comparison against telemetry from an actual operating constellation would test whether the orbital-mechanics coupling remains dominant in practice.

Load-bearing premise

The coupled indirect and direct Markov chains produce a periodic steady state over the right ascension of the ascending node cycle that can be obtained via fixed-point iteration.

What would settle it

Monte Carlo simulation of the full stochastic process with realistic orbital mechanics and failure rates would falsify the claim if the long-run empirical distributions deviate materially from the predicted stationary distributions or if the fixed-point iteration fails to converge.

Figures

Figures reproduced from arXiv: 2605.20875 by Koki Ho, Seungyeop Han, Shoji Yoshikawa, Takumi Noro, Takumi Suda.

Figure 1
Figure 1. Figure 1: illustrates the hybrid strategy. When a satellite fails, an in￾plane spare is used for immediate replacement. If the number of in-plane spares falls below a specified threshold, the policy can use either the indirect channel, by transferring spares from a parking orbit, or the direct channel, by launching new spares directly to the in-plane orbit. At the same time, if the parking orbit stock drops below th… view at source ↗
Figure 2
Figure 2. Figure 2: Stock level profile of (a) constellation and (b) parking orbits under hybrid [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Illustration of Two Approaches 3.2. Constellation Orbit Analysis Method In this section, we introduce the method for computing the periodic sta￾tionary distribution of in-plane inventory. First, the maximum number of satellites in a constellation orbit is Nsatc = max {rc,d + qc,d, rc,i + qc,i + qc,d} , in units of satellites. This bound captures the worst-case scenario in which a direct batch arrives immed… view at source ↗
Figure 4
Figure 4. Figure 4: In-plane state transitions under hybrid replenishment, showing IO/LT phases [PITH_FULL_IMAGE:figures/full_fig_p019_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Comparison results for the representative case of large error: (a) [PITH_FULL_IMAGE:figures/full_fig_p042_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Parking-spare demand distribution from in-plane orbits for the representative [PITH_FULL_IMAGE:figures/full_fig_p044_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Time profile of selected in-plane state probabilities for the representative case [PITH_FULL_IMAGE:figures/full_fig_p045_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Total cost variation of each method at different [PITH_FULL_IMAGE:figures/full_fig_p049_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Relative channel usage in Hybrid Strategy at different [PITH_FULL_IMAGE:figures/full_fig_p050_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Total cost variation of each method at different [PITH_FULL_IMAGE:figures/full_fig_p050_10.png] view at source ↗
read the original abstract

This paper presents a Markov-chain-based method for the early-phase analysis and design of hybrid spare-management architectures for large-scale satellite constellations.} The hybrid strategy combines two channels: an indirect path that stages spares in parking orbits via heavy launch for later transfer to constellation planes, and a direct path that delivers spares to in-plane orbits using small launch vehicles. {To assess the long-run viability of such concepts of operations, satellite failure and replenishment processes are modeled as a Markov chain:} the indirect channel follows a periodic-review reorder-point/order-quantity policy, while the direct channel uses a standard reorder-point/order-quantity policy. These coupled chains yield a periodic steady state over the right ascension of the ascending node cycle via fixed-point iteration, and the stationary distributions provide rigorous cost and resilience metrics. By directly modeling the stochastic, multi-echelon dynamics governed by orbital mechanics, our framework avoids the aggregation assumptions of prior works and remains valid across a wider operating domain. We also introduce an approximate analysis that preserves delay statistics while significantly reducing model size. Building on this fast, accurate analysis, we formulate a cost minimization problem with resilience constraints and solve it using a genetic algorithm. The framework is channel-neutral; the optimization autonomously selects the preferred path and roles. {A case study validates the analysis against Monte Carlo simulations and demonstrates the practical value of the framework in identifying the conditions under which the hybrid policy outperforms pure strategies.

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 / 2 minor

Summary. The paper presents a Markov-chain-based method for early-phase analysis and design of hybrid spare-management architectures for large-scale satellite constellations. The hybrid strategy combines indirect (periodic-review reorder-point/order-quantity) and direct ((r,Q)) resupply channels. Satellite failure and replenishment are modeled as coupled Markov chains that yield a periodic steady state over the right ascension of the ascending node cycle via fixed-point iteration; stationary distributions supply cost and resilience metrics. An approximate analysis that preserves delay statistics is introduced to reduce model size. A cost-minimization problem with resilience constraints is solved via genetic algorithm. The framework is channel-neutral and is validated against Monte Carlo simulations in a case study.

Significance. If the fixed-point iteration converges reliably, the work supplies a granular stochastic model of multi-echelon dynamics that incorporates orbital mechanics without the aggregation assumptions common in prior literature, thereby supporting wider operating domains and autonomous selection of resupply paths. The Monte Carlo validation and the genetic-algorithm optimization that treats the two channels symmetrically constitute concrete strengths that enhance practical applicability for constellation design.

major comments (2)
  1. [Modeling section] Modeling section (paragraph describing fixed-point iteration over the RAAN cycle): the procedure for computing the periodic steady state of the coupled indirect and direct Markov chains is stated without a convergence proof, contraction-mapping argument, monotonicity guarantee, or spectral-radius bound on the iteration operator. Because the cost and resilience metrics are direct functions of this stationary distribution and the subsequent genetic-algorithm optimization relies on it, the absence of such guarantees is load-bearing for the central claim of reliable analysis across hybrid regimes.
  2. [Case-study validation] Case-study validation paragraph: the claim of validation against Monte Carlo simulations is made without quantitative error metrics (e.g., maximum or average relative error in stationary probabilities, cost estimates, or delay statistics) or explicit discussion of how orbital-mechanics parameters enter the transition probabilities. This weakens the assertion that the framework remains valid across a wider operating domain.
minor comments (2)
  1. [Abstract] The abstract states that the approximate analysis 'preserves delay statistics' but does not identify which statistics are preserved or provide the explicit approximation rule.
  2. [Notation] Notation for the state spaces of the two coupled chains and for the RAAN-cycle indexing would benefit from a compact summary table.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for the constructive feedback, which helps clarify the theoretical and empirical foundations of our Markov-chain framework. We respond to each major comment below, indicating planned revisions where appropriate.

read point-by-point responses

  1. Referee: [Modeling section] Modeling section (paragraph describing fixed-point iteration over the RAAN cycle): the procedure for computing the periodic steady state of the coupled indirect and direct Markov chains is stated without a convergence proof, contraction-mapping argument, monotonicity guarantee, or spectral-radius bound on the iteration operator. Because the cost and resilience metrics are direct functions of this stationary distribution and the subsequent genetic-algorithm optimization relies on it, the absence of such guarantees is load-bearing for the central claim of reliable analysis across hybrid regimes.

    Authors: We agree that the manuscript would benefit from additional discussion of convergence. Establishing a general contraction-mapping or spectral-radius bound for the coupled periodic system is non-trivial given the orbital periodicity and channel coupling, and we do not claim such a proof in the current work. In the revision we will add numerical evidence of convergence (iteration counts, residual norms, and success rates) across the parameter ranges used in the case study and optimization, together with a heuristic argument based on the contractive nature of the individual channel transition matrices under standard failure-rate assumptions. revision: partial

  2. Referee: [Case-study validation] Case-study validation paragraph: the claim of validation against Monte Carlo simulations is made without quantitative error metrics (e.g., maximum or average relative error in stationary probabilities, cost estimates, or delay statistics) or explicit discussion of how orbital-mechanics parameters enter the transition probabilities. This weakens the assertion that the framework remains valid across a wider operating domain.

    Authors: We accept this observation. The revised manuscript will report explicit quantitative metrics, including maximum and average relative errors between the Markov-model stationary probabilities, cost values, and delay statistics versus Monte Carlo runs. We will also expand the modeling section to detail how orbital-mechanics parameters (RAAN cycle length, transfer times, and plane-specific access windows) are encoded in the transition probabilities. revision: yes

standing simulated objections not resolved
  • Formal convergence proof, contraction-mapping argument, or spectral-radius bound for the fixed-point iteration

Circularity Check

0 steps flagged

No significant circularity; model-derived metrics are self-contained and externally validated

full rationale

The paper defines a Markov chain for the coupled indirect (periodic-review) and direct (r,Q) replenishment processes, obtains the periodic steady state over the RAAN cycle via fixed-point iteration on the transition structure, and computes cost/resilience metrics directly from the resulting stationary distributions. This is a standard forward derivation from the model equations rather than any reduction by construction; the stationary distribution is solved from the defined transition matrix, not fitted or renamed from external data. The case study explicitly validates the analysis against Monte Carlo simulations, providing an independent external benchmark. No load-bearing self-citations, no fitted parameters presented as predictions, and no uniqueness theorems imported from prior author work appear in the derivation chain. The claim of avoiding aggregation assumptions is an independent modeling choice supported by the direct stochastic formulation and simulation checks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard Markov chain theory plus domain assumptions about orbital periodicity and inventory policies; no new physical entities are introduced.

axioms (2)
  • domain assumption Satellite failure and replenishment processes can be represented as a discrete-time Markov chain with the stated (s,Q) policies.
    Invoked in the modeling paragraph of the abstract.
  • domain assumption The coupled chains admit a periodic steady-state distribution over the RAAN cycle that is reachable by fixed-point iteration.
    Stated as yielding the stationary distributions used for cost and resilience metrics.

pith-pipeline@v0.9.0 · 5797 in / 1447 out tokens · 35777 ms · 2026-05-21T03:48:27.802223+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

36 extracted references · 36 canonical work pages · 2 internal anchors

  1. [1]

    Poghosyan, I

    A. Poghosyan, I. Lluch, H. Matevosyan, A. Lamb, C. Moreno, C. Taylor, A. Golkar, J. Cote, S. Mathieu, S. Pierotti, et al., Unified classification for distributed satellite systems, in: Proceedings of the 4th International Federated and Fractionated Satellite Systems Workshop, Rome, Italy, 2016

    work page 2016

  2. [2]

    Bouwmeester, A

    J. Bouwmeester, A. Menicucci, E. Gill, Improving cubesat reliability: Subsystem redundancy or improved testing?, Reliability Engineering & System Safety 220 (2022) 108288.doi:10.1016/j.ress.2021.108288

    work page doi:10.1016/j.ress.2021.108288 2022

  3. [3]

    Cornara, T

    S. Cornara, T. Beech, M. Belló-Mora, A. Martinez de Aragon, Satellite constellation launch, deployment, replacement and end-of-life strategies, in: 13th Annual AIAA/USU Conference on Small Satellites, 1999

    work page 1999

  4. [4]

    M. A. Luu, D. E. Hastings, On-orbit servicing system architectures for proliferated low-earth-orbit constellations, Journal of Spacecraft and Rockets 59 (6) (2022) 1946–1965.doi:10.2514/1.A35393. 52

    work page doi:10.2514/1.a35393 2022

  5. [5]

    American Enterprise Institute, Moore’s law meets musk’s law: The un- derappreciated story of spacex and the stunning decline in launch costs, https://www.aei.org/articles/moores-law-meet-musks-law-the -underappreciated-story-of-spacex-and-the-stunning-decline -in-launch-costs/, accessed: 2025-07-14 (2024)

    work page 2025

  6. [6]

    here’s how to watch,https://www.space.com/arianespace-o neweb-5-satellites-launch-webcast, accessed: 2025-07-14 (2021)

    Space.com, Arianespace to launch new fleet of oneweb internet satellites tonight. here’s how to watch,https://www.space.com/arianespace-o neweb-5-satellites-launch-webcast, accessed: 2025-07-14 (2021)

    work page 2025

  7. [7]

    McDowell, Starlink statistics,https://planet4589.org/space/con /star/stats.html, accessed: 2025-07-14 (2025)

    J. McDowell, Starlink statistics,https://planet4589.org/space/con /star/stats.html, accessed: 2025-07-14 (2025)

    work page 2025

  8. [8]

    Space.com, Starlink satellites: Facts, tracking and impact on astron- omy,https://www.space.com/spacex-starlink-satellites.html, accessed: 2025-08-07 (2025)

    work page 2025

  9. [9]

    CNBC, Amazon launches second batch of kuiper internet satellites, tak- ing on elon musk’s starlink,https://www.cnbc.com/2025/06/23/amaz on-kuiper-satellites-musk-starlink.html, accessed: 2025-07-14 (2025)

    work page 2025

  10. [10]

    Telecoms Tech News, China launches first satellites for guowang con- stellation,https://www.telecomstechnews.com/news/china-launc hes-first-satellites-guowang-constellation/, accessed: 2025- 07-14 (2024)

    work page 2025

  11. [11]

    An Updated Comparison of Four Low Earth Orbit Satellite Constellation Systems to Provide Global Broadband,

    N. Pachler, I. del Portillo, E. F. Crawley, B. G. Cameron, An up- dated comparison of four low earth orbit satellite constellation sys- 53 tems to provide global broadband, in: 2021 IEEE International Confer- ence on Communications Workshops (ICC Workshops), 2021, pp. 1–7. doi:10.1109/ICCWorkshops50388.2021.9473799

    work page doi:10.1109/iccworkshops50388.2021.9473799 2021

  12. [12]

    Pachler, E

    N. Pachler, E. F. Crawley, B. G. Cameron, Flooding the market: Com- paring the performance of nine broadband megaconstellations, IEEE Wireless Communications Letters 13 (9) (2024) 2397–2401.doi: 10.1109/LWC.2024.3416531

    work page doi:10.1109/lwc.2024.3416531 2024

  13. [13]

    Dishon, G

    M. Dishon, G. H. Weiss, A communications satellite replenishment pol- icy, Technometrics 8 (3) (1966) 399–409.doi:10.1080/00401706.196 6.10490373

    work page doi:10.1080/00401706.196 1966

  14. [14]

    B. R. Sumter, Optimal replacement policies for satellite constellations, Master’s thesis, Air Force Institute of Technology (2003)

    work page 2003

  15. [15]

    Jakob, S

    P. Jakob, S. Shimizu, S. Yoshikawa, K. Ho, Optimal satellite constel- lation spare strategy using multi-echelon inventory control, Journal of Spacecraft and Rockets 56 (5) (2019) 1449–1461.doi:10.2514/1.A343 87

    work page doi:10.2514/1.a343 2019

  16. [16]

    R. V. Ramasesh, J. K. Ord, J. C. Hayya, A. Pan, Sole versus dual sourcing in stochastic lead-time (s, q) inventory models, Management science 37 (4) (1991) 428–443.doi:10.1287/mnsc.37.4.428

    work page doi:10.1287/mnsc.37.4.428 1991

  17. [17]

    Chiang, W

    C. Chiang, W. Benton, Sole sourcing versus dual sourcing under stochas- tic demands and lead times, Naval Research Logistics (NRL) 41 (5) (1994) 609–624.doi:10.1002/1520- 6750(199408)41:5<609:: AID-NAV3220410503>3.0.CO;2-7. 54

    work page doi:10.1002/1520- 1994

  18. [18]

    A. C. Pan, R. V. Ramasesh, J. C. Hayya, J. K. Ord, Multiple sourcing: the determination of lead times, Operations Research Letters 10 (1) (1991) 1–7.doi:10.1016/0167-6377(91)90079-5

    work page doi:10.1016/0167-6377(91)90079-5 1991

  19. [19]

    Schimpel, Dual sourcing : with arbitrary stochastic demand and stochastic lead times, KIT Scientific Publishing, 2010.doi:10.5445/ KSP/1000018435

    U. Schimpel, Dual sourcing : with arbitrary stochastic demand and stochastic lead times, KIT Scientific Publishing, 2010.doi:10.5445/ KSP/1000018435

    work page arXiv 2010

  20. [20]

    T. Sung, J. Ahn, Optimal deployment of satellite mega-constellation, Acta Astronautica 202 (2023) 653–669.doi:10.1016/j.actaastro.20 22.10.027

    work page doi:10.1016/j.actaastro.20 2023

  21. [21]

    J. Kim, J. Ahn, T. Sung, Replenishment strategy for satellite constella- tion with dual supply modes, Journal of Spacecraft and Rockets 62 (5) (2025) 1567–1583.doi:10.2514/1.A36281

    work page doi:10.2514/1.a36281 2025

  22. [22]

    S. Han, T. Noro, K. Ho, Analysis and design of satellite constellation spare strategy using markov chain, in: 2024 AAS/AIAA Astrodynamics Specialist Conference, 2024, broomfield, Colorado, USA, August 2024

    work page 2024

  23. [23]

    S. Han, Z. Grieser, S. Yoshikawa, T. Noro, T. Suda, K. Ho, Analysis and design of spare strategy for large-scale satellite constellation using direct insertion under(r, q)policy, arXiv preprint arXiv:2509.10585 (2025). doi:10.48550/arXiv.2509.10585

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2509.10585 2025

  24. [24]

    S. Han, Z. Grieser, S. Yoshikawa, T. Noro, T. Suda, K. Ho, Spare strat- egy analysis and design for mega satellite constellations using Markov chain, Journal of Spacecraft and Rockets, accepted for publication (2026).arXiv:2509.09957. 55

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  25. [25]

    J. G. Walker, Satellite constellations, Journal of the British Interplane- tary Society 37 (1984) 559

    work page 1984

  26. [26]

    J. E. Prussing, B. A. Conway, Orbital mechanics, Oxford University Press, USA, 1993

    work page 1993

  27. [27]

    F. S. Hillier, G. J. Lieberman, Introduction to operations research, McGraw-Hill, 2015

    work page 2015

  28. [28]

    Greenbaum, Iterative Methods for Solving Linear Systems, SIAM, 1997.doi:10.1137/1.9781611970937

    A. Greenbaum, Iterative Methods for Solving Linear Systems, SIAM, 1997.doi:10.1137/1.9781611970937

    work page doi:10.1137/1.9781611970937 1997

  29. [29]

    Department of Defense, Fact sheet: Resilience of space capabilities (2011)

    U.S. Department of Defense, Fact sheet: Resilience of space capabilities (2011)

    work page 2011

  30. [30]

    1002--1007

    R. Burch, A method for calculation of the resilience of a space system, in: MILCOM 2013 - 2013 IEEE Military Communications Conference, 2013, pp. 1002–1007.doi:10.1109/MILCOM.2013.174

    work page doi:10.1109/milcom.2013.174 2013

  31. [31]

    SpaceInsider, How much does it cost to launch a rocket? [by type & size],https://spaceinsider.tech/2023/08/16/how-much-does-i t-cost-to-launch-a-rocket/, accessed: 2025-07-11 (2025)

    work page 2023

  32. [32]

    Ainvest, Rocket lab’s 48-hour launch turnaround: A strategic play for dominance in the smallsat launch market,https://www.ainvest.co m/news/rocket-lab-48-hour-launch-turnaround-strategic-pla y-dominance-smallsat-launch-market-2506/, accessed: 2025-07-11 (2025). 56

    work page 2025

  33. [33]

    Spaceflight Now, Spacex breaks launchpad turnaround record with mid- night starlink flight,https://spaceflightnow.com/2025/06/27/live -coverage-spacex-to-launch-27-starlink-satellites-on-falco n-9-rocket-from-cape-canaveral-4/, accessed: 2025-07-11 (2025)

    work page 2025

  34. [34]

    G. F. Dubos, J.-F. Castet, J. H. Saleh, Statistical reliability analysis of satellites by mass category: Does spacecraft size matter?, Acta Astro- nautica 67 (5) (2010) 584–595.doi:10.1016/j.actaastro.2010.04. 017

    work page doi:10.1016/j.actaastro.2010.04 2010

  35. [35]

    Langer, Reliability assessment and reliability prediction of cubesats through system level testing and reliability growth modelling, Ph.D

    M. Langer, Reliability assessment and reliability prediction of cubesats through system level testing and reliability growth modelling, Ph.D. thesis, Technische Universität München (2018)

    work page 2018

  36. [36]

    S. A. Jacklin, Small-satellite mission failure rates, NASA Technical Memorandum NASA/TM-2018-220034, NASA Ames Research Center, Moffett Field, CA (March 2019). 57

    work page 2018