Pith. sign in

REVIEW 3 major objections 5 minor 33 references

Meta-heuristic design of a light-weight homologous backup structure of the primary reflector for the Large Submillimeter Telescope

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A genetic-algorithm search over truss geometry and member sizes finds a 50 m-class telescope backup structure that keeps the primary reflector at about 5 µm RMS surface accuracy under gravity with only a small number of actively…

desk verdict Solid GA-based BUS optimization let down by an abstract that sells a hub-excluded actuator-stroke RMS as a full-aperture surface error. read the letter →

arxiv 2506.02191 v1 pith:LM7QMUQV submitted 2025-06-02 astro-ph.IM

Chihiro Imamura , Yoichi Tamura , Hiroaki Kawamura , Toshiaki Kimura , Akio Taniguchi , Mikio Kurita This is my paper
classification astro-ph.IM
keywords geneticalgorithmstructuraloptimizationhomologousdeformationbackupstructuresubmillimetertelescopetrussactuatorstrokesurfaceaccuracy
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tries to show that a genetic-algorithm search can design the backup structure of a 50 m-class submillimeter telescope to be both light and precise enough for the next generation of instruments. The authors model the backup structure as a steel truss, group its design variables, and optimize simultaneously for low mass and for small maximum actuator stroke under gravity, with constraints on aperture efficiency and member stress. The central result is an axisymmetric truss that keeps the primary reflector at roughly $5\,\mu\mathrm{m}$ RMS surface error while only a small number of nodes near the central hub need active control. This matters because proposed telescopes such as the Large Submillimeter Telescope need about $45\,\mu\mathrm{m}$ RMS or better while remaining cheap and light enough to build and steer.

What carries the argument

The carrying mechanism is homologous deformation — a structure whose gravity-deformed shape remains close to a paraboloid so that a refocus or small corrections restore the surface — searched by an NSGA-II genetic algorithm over grouped truss variables: radial and vertical nodal displacements chosen from discrete tables, plus cross-sectional areas of straight elements. The objective rewards small maximum actuator stroke, averaged over elevations of 85° and 30° after adjusting the surface at 50°, which is a proxy for how well the deformed shape matches the ideal paraboloid. Penalty terms enforce aperture efficiency above 90% and allowable stress computed from a steel design code. The decisive device is the central-hub treatment: nodes on the hub are excluded from the accuracy evaluation during optimization and then corrected separately, which is what makes the roughly $5\,\mu\mathrm{m}$ figure reachable.

What would settle it

Re-run the optimized truss in a full finite-element model with the real Ritchey-Chrétien hyperboloid, the actual segmented-panel masses, secondary-mirror loads, and wind/thermal gradients at elevations between 30° and 85°, then recompute RMS including the hub nodes and their roughly $300\,\mu\mathrm{m}$ actuator corrections; if the result exceeds about $5\,\mu\mathrm{m}$ RMS (or the $45\,\mu\mathrm{m}$ specification), the central claim is not transferable.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that minimizing the maximum actuator stroke length — rather than surface error itself — is an effective surrogate for reflector accuracy, and that this surrogate lets a multi-objective genetic algorithm find a 50 m-class truss whose surface stays within about $5\,\mu\mathrm{m}$ RMS under gravitational loading. The best solution is the axisymmetric model optimized with the central hub treated as rigid and hub nodes excluded from the accuracy metric; those excluded nodes then need roughly $300\,\mu\mathrm{m}$ of actuator stroke to be pulled back to the ideal surface. Every optimized structure satisfies the 90% aperture-efficiency constraint and the steel allowable-stress limits, and the solutions trace a clear mass-versus-accuracy trade-off. The non-axisymmetric variants do not beat the axisymmetric one within the 100,000-generation budget, which the authors attribute to their much larger search space.

Load-bearing premise

The $\sim 5\,\mu\mathrm{m}$ figure assumes gravity is the only force on the structure, the reflector is a simple paraboloid rather than its real hyperboloid shape, the surface load is a single uniform $55\,\mathrm{kg}/\mathrm{m}^2$ value, and the central hub is left out of the accuracy count; if wind, temperature, or secondary-mirror forces matter, the number will not transfer to the built telescope.

Editorial extensions

If this is right

  • A 50 m-class submillimeter dish can meet its surface-accuracy requirement with a small number of active actuators rather than a fully active surface, reducing cost and complexity.
  • The strong correlation between maximum actuator stroke and RMS surface error means future optimizations can keep using stroke length as a cheap, practical objective.
  • Axisymmetric trusses remain competitive with non-axisymmetric ones; the extra freedom of asymmetric designs did not pay off within the search budget, so more iterations or better seeding are needed before ruling them out.
  • Hub-excluded accuracy comes at a predictable price: actuators near the central hub need strokes of roughly 200–300 µm, so maintenance and wear will concentrate there.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the $5\,\mu\mathrm{m}$ RMS is computed under gravity-only loading on a uniform $55\,\mathrm{kg}/\mathrm{m}^2$ load and a paraboloid rather than the true Ritchey-Chrétien hyperboloid, a fair next test is to re-run the best solution under wind, thermal, and secondary-mirror loads; the figure is best read as a lower-bound estimate of achievable precision.
  • The non-axisymmetric search might be greatly improved by seeding its initial population with the best axisymmetric solution, a step the paper suggests but does not execute; the asymmetric freedom may then start to pay off.
  • The actuator-stroke surrogate could lose its fidelity once dynamics enter: under wind gusts the highest-stroke node may not be the one that most degrades the beam, so adding a natural-frequency or dynamic-response objective would likely shift the Pareto front.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper applies a multi-objective genetic algorithm (NSGA-II) with OpenSeesPy finite-element evaluation to optimize the shape and sizing of a 50 m-class backup structure (BUS) for the Large Submillimeter Telescope. The design variables are grouped nodal positions and truss cross-sectional areas; the two objectives are maximum actuator stroke demand and BUS mass, subject to constraints on aperture efficiency (via focal-length offset) and buckling stress. Four optimization cases are considered, combining axisymmetric/non-axisymmetric geometry with and without a hub-rigidity treatment that excludes central-hub nodes from the objective and accuracy evaluation. The best reported result is 4.799 µm RMS in the axisymmetric hub-rigidity case, with hub nodes requiring ~300 µm strokes. The paper also reports first natural frequencies and compares actuator-stroke demand with surface accuracy.

Significance. If the headline result is properly qualified, the paper is a useful demonstration of GA-based homologous design at the 50 m scale. The aperture-efficiency derivation (Eqs. 4–11), the stress constraints (Eqs. 13–16), and the use of a standard FE solver are coherent, and the four-case comparison with natural-frequency follow-up gives the study practical engineering content. However, the central claim as stated in the Abstract and Conclusion overreaches the computed metric: the ~5 µm RMS value is a hub-excluded RMS of required actuator strokes, not a full-aperture optical surface error, and it is obtained from the same simplified FE model used for the optimization objective. These qualifications are present in the body but absent from the headline, so the manuscript needs substantial revision of its claims.

major comments (3)
  1. [Abstract; §2.1.4; §3.3; Eqs. (22)–(23)] The headline "~5 µm RMS" is not a full-aperture optical surface error. The value 4.799 µm reported in §3.3 is an RMS of required actuator stroke demand (δEL,i − δ50,i) computed from Eq. (22) after explicitly ignoring the central-hub nodes as described in §2.1.4; §3.3 itself states that the excluded hub node requires about 300 µm of stroke (Fig. 18). The Abstract and Conclusion present the 5 µm figure without this caveat, so the central claim overstates what is computed. Please report hub-included values of ε85, ε30, and εeval, the number and area fraction of hub nodes, and either quantify the "tiny portion" of actively controlled nodes or rephrase the claim as a hub-excluded stroke-demand RMS.
  2. [§2.1.3, Eqs. (1)–(2) and (17)–(23); Fig. 23] The objective D and the reported accuracy εeval are both derived from the same deformation field and the same ideal-surface fit with free focal parameters. Optimizing the maximum actuator stroke is therefore essentially optimizing a quantity tightly correlated with εeval, so the strong correlation shown in Fig. 23 is partly by construction rather than an independent validation of εeval as a predictor of optical performance. In addition, εeval is a required-stroke metric, not the residual surface error after actuator correction; the achieved accuracy also depends on actuator resolution and control errors, which are not modeled. Please state explicitly that εeval is a lower-bound stroke-demand proxy, and if feasible include a full-aperture residual RMS after simulated actuator correction.
  3. [§2.1.1.1, §2.2.1, Table 4] The numerical demonstration is restricted to gravitational loading on a paraboloidal BUS model with a uniform 55 kg/m² blanket load, with wind, thermal, and secondary-mirror loads explicitly excluded. These limitations are stated in §2.1.1.1 and Table 4, and they are appropriate for a first demonstration. However, the Abstract and Conclusion draw the stronger conclusion that the method can meet the stringent surface-accuracy requirements of next-generation telescopes. Please qualify the claim as a gravity-only, paraboloid-model, hub-excluded result and add a sentence discussing the expected impact of omitted load cases on the achievable full-aperture accuracy.
minor comments (5)
  1. [§4.1, §3.4, Table 1] There are several small language and typographical errors: "asynmetic bowing" in §4.1, "figure21" without a space in §3.4, and "secodary reflecotor" in Table 1.
  2. [Table 2] The allele table has inconsistent spacing and plus signs (e.g., "+0 .105", "+0 .225", "1 .425"), which makes the discrete design-variable values harder to read; please retypeset uniformly.
  3. [§2.1.3.4 and §2.1.4] Eq. (22) states that hub nodes are ignored in εθ before the hub-rigidity method is introduced in §2.1.4; reorder the sections or add a forward reference to avoid confusing the reader.
  4. [Figs. 22–23 and captions] The legend text in Fig. 22/23 is awkward ("Open circle represent..." and overlapping uses of gray/yellow and filled/open markers); clarify the marker conventions and correct the duplicated "Alt text" in the Fig. 17 caption.
  5. [Code and data availability] The paper uses OpenSeesPy and DEAP but provides no code or data availability statement; a reproducibility section with the model geometry, grouping definitions, and GA settings would strengthen the paper.

Circularity Check

0 steps flagged · score 0.0 of 10

The reported 5 µm RMS figure is an optimized performance metric from the paper's own finite-element model, not a fitted input recycled as an external prediction, and the derivation is self-contained.

full rationale

The paper is a numerical structural optimization study rather than an empirical prediction exercise, so the circularity patterns do not apply in a load-bearing way. The objective functions are the maximum actuator stroke D (Eq. 2) and BUS mass W (Eq. 3); the reported surface accuracy εθ (Eq. 22) and εeval (Eq. 23) are computed from the same finite-element displacement field as Dθ (Eq. 1). This makes the ~5 µm RMS an optimized performance metric of the designed structure, not a parameter fitted to a subset of data and then presented as a prediction of a closely related quantity; no external dataset is claimed to be forecast. The ideal-surface fit with free focal parameters is standard homologous-deformation practice, and the hub-node exclusion is explicitly disclosed in §2.1.4 and §3.3 rather than hidden. Self-citations to Kurita et al. (2010) and Shintani et al. (2024) are contextual or offered as future suggestions, not used to justify the present numerical results, which are reproduced in the paper via OpenSeesPy and DEAP. The abstract's '~5 µm RMS' claim would be clearer if it carried the hub-exclusion caveat, but that is a clarity and overstatement concern, not a circular derivation step.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The central design rests on standard truss mechanics plus several simplifying assumptions: gravity-only loading, a paraboloid reflector, a fixed lumped load, hand-chosen GA settings, and visually assessed convergence. No new physical entities are introduced; the only 'fitted' element is the standard least-squares ideal-surface fit used to report homologous residual errors.

free parameters (6)
  • Constraint safety factor = 0.95
    Penalty terms activate at 95% of the calculated aperture-efficiency and stress limits; the 5% margin is chosen by hand and affects which designs survive.
  • Elevation angles used in objective = 85° and 30° with 50° reference
    The actuator-stroke objective averages the maximum deviation at ELs of 85° and 30°, an arbitrary bracket of the operating range.
  • Surface load per area = 55 kg/m²
    Lumped from components in Table 4; a different load model changes the deformation and the optimized structure.
  • GA hyperparameters = population 300, crossover 0.7, mutation 0.01, 100,000 generations
    NSGA-II settings are fixed without a sensitivity study; the reported optima are from a single run.
  • Design variable grouping = 5 node groups, 60 cross-section groups
    Grouping reduces the search space from ~10^4800 to ~10^70 and constrains the achievable geometries.
  • Ideal surface fit degrees of freedom = focal point and vertex coordinates
    The reported surface error is the residual after least-squares fitting the ideal paraboloid, which removes the best-fit focus and vertex shifts.
assumptions (6)
  • domain assumption Only gravitational loads act on the BUS.
    Section 2.1.1.1 states external disturbances, gusts, and secondary reflector loads are out of scope; surface accuracy results depend on this loading model.
  • domain assumption The primary reflector is modeled as a paraboloid rather than the hyperboloid of the actual Ritchey-Chrétien optics.
    Stated in Section 2.2.1; this simplifies the optical evaluation and may change the optimal structure.
  • standard math Truss elements are pin-jointed and carry only axial stress; buckling is checked with the allowable stress design code.
    Standard structural engineering idealization, stated in Section 2.1.3.2.
  • domain assumption Aperture efficiency constraint is derived assuming rigid segmented panels, uniform illumination, and λ=350 µm.
    Section 2.1.3.2 derives the 70 mm focal-length-offset limit from these assumptions; different assumptions change the constraint.
  • ad hoc to paper The genetic algorithm converges to near-optimal solutions within 100,000 generations.
    Convergence is asserted from visual inspection of Figures 8, 13, 16, and 19, not from a formal stopping criterion or repeated runs.
  • standard math The finite element solver (OpenSeesPy) provides a faithful model of the truss deformation.
    Trust in the structural solver is assumed; no experimental validation is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Meta-heuristic design of a light-weight homologous backup structure of the primary reflector for the Large Submillimeter Telescope." pith.science (2026). https://pith.science/paper/LM7QMUQV

@misc{pith2026250602191,
  author       = {Pith},
  title        = {Pith review of: Meta-heuristic design of a light-weight homologous backup structure of the primary reflector for the Large Submillimeter Telescope},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LM7QMUQV}},
  note         = {Machine review of arXiv:2506.02191}
}
abstract

The development of large-aperture submillimeter telescopes, such as the Large Submillimeter Telescope (LST) and the Atacama Large Aperture Submillimeter Telescope (AtLAST), is essential to overcome the limitations of current observational capabilities in submillimeter astronomy. These telescopes face challenges related to maintaining high surface accuracy of the main reflector while minimizing the weight of the telescope structure. This study introduces a genetic algorithm (GA)-based structural optimization, previously applied in related works, to 50 m-class backup structures (BUSes) with a variable focal position, addressing the challenge of achieving both lightweight construction and high surface accuracy through the consideration of homologous deformation. We model the BUS as a truss structure and perform multi-objective optimization using a GA. The optimization process considers two structures: axisymmetric and non-axisymmetric between the top and bottom. The optimization aims to find structures that simultaneously minimize the maximum stroke length of actuators and the mass of the BUS under practical constraints. The optimized structures show improved surface accuracy, primarily due to the minimization of the maximum actuator stroke length, and reduced weight, both achieved under the imposed constraints. Notably, we find a homologous BUS solution that achieves a surface error of down to $\sim 5\,\mu\mathrm{m}$ RMS with a tiny portion of the truss nodes being actively controlled. The results highlight the potential of GA-based optimization in the design of next-generation submillimeter telescopes, suggesting that further exploration of non-axisymmetric structures could yield even more effective solutions. Our findings support the application of advanced optimization techniques to achieve high-performance and cost-effective telescope designs.

Figures

Figures reproduced from arXiv: 2506.02191 by the authors.

Figure 1
Figure 1. Schematic diagram of the BUS cross-section. The BUS is com￾posed of nodes and straight elements and is supported by the central hub. Alt text: Cross-sectionnal view of the BUS in this study. area. The relation between these design variables and the allele number, a number that describes a variable in GA, in our optimiza￾tion process is summarized in table 2. We note that the gray nodes in figures 2 and 3 are elimina… view at source ↗
Figure 2
Figure 2. Initial nodal positions and straight truss elements of the axial symmetric structure in the three view drawings. The origin is set at the bottom of the main reflector and the axes are fixed to the reflector. The reflector rotates on the ξ-axis. Colored groups of the truss component in each figure vary its position or cross-sectional area following the corresponding allele. We note that gray points in (a), (c), and (… view at source ↗
Figure 3
Figure 3. Initial nodal positions and straight truss elements of the non-axisymmetric structure in the three view drawings. The origin is set at the bottom of the main reflector and the axes are set right-handed and fixed to the reflector. The reflector rotates on the ξ-axis. Colored groups of the truss component in each figure vary its position or cross-sectional area following the corresponding allele. We note that gray poi… view at source ↗
Figures from the paper (20 more)
Figure 4
Figure 4. Figure 4: Flow chart of the structural optimization. Alt text: Flow chart. repeats, returning to the evaluation step if the maximum number of generations has not been reached; otherwise, the resulting genes represent the optimal solution. We also implement the concept of a “hall…
Figure 6
Figure 6. Figure 6: Aperture efficiency as a function of focal length offset at observing wavelength λ=350 µm assuming a uniform illumination pattern (w=1) and an aperture of 50 m. The solid blue line corresponds to a segmented mirror edge length of 1 m, the dashed orange line to 2 m, and…
Figure 5
Figure 5. Figure 5: (a) Schematic diagram of distribution of surface error due to dis￾crepancy between ideal surface and gravitationally deformed surface. ξ ′η ′ ζ ′ -coordinate system in this figure has an origin on the center of the panel, with the ξ ′η ′ -plane parallel to the tangent …
Figure 7
Figure 7. Figure 7: Conceptual diagram of the evaluation of structural optimization con￾sidering the rigidity of the central hub. The rigid hub supports the de￾formed BUS (dashed gray line), which may hinder homologous deforma￾tion. Evaluating the objective function without surface nodes …
Figure 9
Figure 9. Figure 9: Randomly generated structures and optimized structures in the surface accuracy and mass of the structure plane in the fiducial case. Blue circles represent randomly generated initial generation and solid orange circles show final generation. Alt text: Scatter graph com…
Figure 10
Figure 10. Figure 10: Distribution of actuator stroke length Dθ of the most rigid structures in the initial and final generations of the fiducial case at ELs of 85◦ and 30◦ , with a common color map scale. Top left: The most rigid structure in the initial generation at EL 85◦ . Top right: …
Figure 11
Figure 11. Figure 11: Cross-sectional view of the optimized structure in the fiducial case at the plane ξ = 0. The width of the elements in this figure corresponds to the actual cross-sectional area of the optimized structure. The straight elements are colored according to their respective…
Figure 12
Figure 12. Figure 12: Maps of actuator stroke length Dθ and the deformed structures of the optimized structure with the most accurate surface in the axisymmetric (fiducial) case. (a) Distribution of actuator stroke lengths at an EL of 85◦ . (b) Deformed BUS at an EL of 85◦ . (c) Distributi…
Figure 13
Figure 13. Figure 13: Statistics in the optimization process of the non-axisymmetric BUS. The upper panel shows the objective function of the maximum actu￾ator stroke, and the lower panel shows the objective function of the mass of the BUS. The red line represents the minimum value of the …
Figure 21
Figure 21. Figure 21: figure 21. Additionally, all optimized structures satisfies the con [PITH_FULL_IMAGE:figures/full_fig_p014_21.png]
Figure 14
Figure 14. Figure 14: Randomly generated structures and optimized structures in the surface accuracy and mass of the structure plane in the non-axisymmetric case. Blue circles represent randomly generated initial generation and solid orange circles show final generation. Alt text: Scatter …
Figure 15
Figure 15. Figure 15: Maps of actuator stroke length Dθ and the deformed structures of the optimized structure with the most accurate surface in the non-axisymmetric case. (a) Distribution of actuator stroke lengths at an EL of 85◦ . (b) Deformed BUS at an EL of 85◦ . (c) Distribution of a…
Figure 16
Figure 16. Figure 16: Statistics in the optimization process of the non-axisymmetric BUS without evaluation of nodes on the central hub. The upper panel shows the objective function of the maximum actuator stroke, and the lower panel shows the objective function of the mass of the BUS. The…
Figure 17
Figure 17. Figure 17: Randomly generated structures and optimized structures in the surface accuracy and mass of the structure plane in the optimization of axisymmetric case considering the hub rigidity. Blue circles represent ran￾domly generated initial generation and solid orange circles…
Figure 18
Figure 18. Figure 18: Maps of actuator stroke length Dθ and the deformed structures of the optimized structure with the most accurate surface in the axisymmetric case considering the hub rigidity. The star in each panel indicates the node on the central hub requiring the maximum actuator s…
Figure 19
Figure 19. Figure 19: Statistics in the optimization process of the axisymmetric BUS without evaluation of nodes on the central hub. The upper panel shows the objective function of the maximum actuator stroke, and the lower panel shows the objective function of the mass of the BUS. The red…
Figure 20
Figure 20. Figure 20: Randomly generated structures and optimized structures in the surface accuracy and mass of the structure plane in the non-axisymmetric case considering the rigidity of the central hub. Blue circles represent randomly generated initial generation and solid orange circl…
Figure 21
Figure 21. Figure 21: Maps of actuator stroke length Dθ and the deformed structures of the optimized structure with the most accurate surface in the non-axisymmetric case considering the hub rigidity. The star in each panel indicates the node on the central hub requiring the maximum actuat…
Figure 22
Figure 22. Figure 22: Comparison of four optimization cases. The gray filled circles rep￾resent the axisymmetric case considering the nodes on the central hub in the optimization, the yellow filled circles represent the non-axisymmetric case considering the nodes on the central hub. Open c…
Figure 23
Figure 23. Figure 23: Comparison between the maximum actuator stroke and surface accuracy. Legend is the same as figure 22. Alt text: Scatter graph [PITH_FULL_IMAGE:figures/full_fig_p020_23.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

33 extracted references · 32 canonical work pages

  1. [1]

    1982, Mitsubishi Denki Giho, 56, 494 Architectural Institute of Japan

    Akabane, K., Katagi, T., Ishii, K., Takisawa, Y ., & Ogata, Y . 1982, Mitsubishi Denki Giho, 56, 494 Architectural Institute of Japan. 2019, AIJ Standard for Allowable Stress Design of Steel Structures (Architectural Institute of Japan)

    work page 1982

  2. [2]

    Baars, J. W. M., Hooghoudt, B. G., Mezger, P. G., & de Jonge, M. J. 1987, A&A, 175, 319

    1987

  3. [3]

    Baars, J. W. M., Martin, R. N., Mangum, J. G., McMullin, J. P., & Peters, W. L. 1999, PASP, 111, 627,

    work page 1999

  4. [4]

    2023, PASP, 135, 095001,

    Chen, M.-T., Asada, K., Matsushita, S., et al. 2023, PASP, 135, 095001,

    work page 2023

  5. [5]

    2002, IEEE Transactions on Evolutionary Computation, 6, 182,

    Deb, K., Pratap, A., Agarwal, S., & Meyarivan, T. 2002, IEEE Transactions on Evolutionary Computation, 6, 182,

    work page 2002

  6. [6]

    2004, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, V ol

    Ezawa, H., Kawabe, R., Kohno, K., & Yamamoto, S. 2004, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, V ol. 5489, Ground-based Telescopes, ed. J. Oschmann, Jacobus M., 763–772

    work page 2004

  7. [7]

    2012, Journal of Machine Learning Research, 13, 2171

    Fortin, F.-A., De Rainville, F.-M., Gardner, M.-A., Parizeau, M., & Gagné, C. 2012, Journal of Machine Learning Research, 13, 2171

    work page 2012

  8. [8]

    The Optical Design Concept for the Atacama Large Aperture Submillimeter Telescope (AtLAST)

    Gallardo, P. A., Puddu, R., Mroczkowski, T., et al. 2024, arXiv e-prints, arXiv:2406.11502,

    work page Pith review arXiv 2024

Show all 33 references
  1. [9]

    2022, Journal of Astronomical Telescopes, Instruments, and Systems, 8, 028001,

    Gao, J., Wang, H., Zuo, Y ., et al. 2022, Journal of Astronomical Telescopes, Instruments, and Systems, 8, 028001,

    2022

  2. [10]

    1992, A&A, 262, 624 Güsten, R., Nyman, L

    Guilloteau, S., Delannoy, J., Downes, D., et al. 1992, A&A, 262, 624 Güsten, R., Nyman, L. Å., Schilke, P., et al. 2006, A&A, 454, L13,

    1992

  3. [11]

    1985, in European Southern Observatory Conference and Workshop

    Hills, R. 1985, in European Southern Observatory Conference and Workshop

    1985

  4. [12]

    Ho, P. T. P., Moran, J. M., & Lo, K. Y . 2004, ApJL, 616, L1,

    2004

  5. [13]

    Holland, J. H. 1975, Adaptation in natural and artificial systems. an intro- ductory analysis with applications to biology, control and artificial intelli- gence

    1975

  6. [14]

    H., Schloerb, F

    Hughes, D. H., Schloerb, F. P., Aretxaga, I., et al. 2020, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, V ol. 11445, Ground-based and Airborne Telescopes VIII, ed. H. K. Marshall, J. Spyromilio, & T. Usuda, 1144522

    2020

  7. [15]

    2016, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, V ol

    Kawabe, R., Kohno, K., Tamura, Y ., et al. 2016, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, V ol. 9906, Ground- based and Airborne Telescopes VI, ed. H. J. Hall, R. Gilmozzi, & H. K. Marshall, 990626

    2016

  8. [16]

    D., Mroczkowski, T

    Klaassen, P. D., Mroczkowski, T. K., Cicone, C., et al. 2020, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, V ol. 11445, Ground-based and Airborne Telescopes VIII, ed. H. K. Marshall, J. Spyromilio, & T. Usuda, 114452F

    2020

  9. [17]

    2010, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, V ol

    Kurita, M., Ohmori, H., Kunda, M., et al. 2010, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, V ol. 7733, Ground- based and Airborne Telescopes III, ed. L. M. Stepp, R. Gilmozzi, & H. J. Hall, 77333E

    2010

  10. [18]

    2020, PASJ, 72, 48,

    Kurita, M., Kino, M., Iwamuro, F., et al. 2020, PASJ, 72, 48,

    2020

  11. [19]

    2020, Appl

    Lou, Z., Zuo, Y .-x., Yao, Q.-j., et al. 2020, Appl. Opt., 59, 3353,

    2020

  12. [20]

    Meinel, A. B. 1982, Journal of the Optical Society of America (1917-1983), 72, 14,

    1982

  13. [21]

    A., Timpe, M., et al

    Mroczkowski, T., Gallardo, P. A., Timpe, M., et al. 2025, A&A, 694, A142,

    2025

  14. [22]

    2022, in Society of Photo- Optical Instrumentation Engineers (SPIE) Conference Series, V ol

    Nakano, S., Tamura, Y ., Taniguchi, A., et al. 2022, in Society of Photo- Optical Instrumentation Engineers (SPIE) Conference Series, V ol. 12185, Adaptive Optics Systems VIII, ed. L. Schreiber, D. Schmidt, & E. Vernet, 121856Z

    2022

  15. [23]

    Phillips, T. G. 2007, The Caltech Submillimeter Observatory (IEEE), 1849–1852

    2007

  16. [24]

    2017, A&A, 608, A40,

    Prandoni, I., Murgia, M., Tarchi, A., et al. 2017, A&A, 608, A40,

    2017

  17. [25]

    A., Mroczkowski, T., et al

    Puddu, R., Gallardo, P. A., Mroczkowski, T., et al. 2024, arXiv e-prints, arXiv:2406.16602,

    2024 arXiv

  18. [26]

    2024, arXiv e-prints, arXiv:2406.08611,

    Reichert, M., Timpe, M., Kaercher, H., et al. 2024, arXiv e-prints, arXiv:2406.08611,

    2024 arXiv

  19. [27]

    2024, Structural and Multidisciplinary Optimization, 67, 64

    Shintani, K., Kawamura, H., Kimura, T., & Yamada, T. 2024, Structural and Multidisciplinary Optimization, 67, 64

    2024

  20. [28]

    2020, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, V ol

    Tamura, Y ., Kawabe, R., Fukasaku, Y ., et al. 2020, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, V ol. 11445, Ground-based and Airborne Telescopes VIII, ed. H. K. Marshall, J. Spyromilio, & T. Usuda, 114451N

    2020

  21. [29]

    Vial, J., Arroyo, D., Canales, C., et al. 2020, The Leighton Chajnantor Telescope: Project update and mechanical structural analysis in prepara- tions for new deployment in Chajnantor, Chile (Society of Photo-Optical Instrumentation Engineers (SPIE)), Art. No. 114453C von Hoer...

    2020

  22. [30]

    2024, Structural and Multidisciplinary Optimization, 67, 5

    Wang, Y ., & Sigmund, O. 2024, Structural and Multidisciplinary Optimization, 67, 5

    2024

  23. [31]

    1970, Nature, 228, 507,

    Wielebinski, R. 1970, Nature, 228, 507,

    1970

  24. [32]

    Wootten, A., & Thompson, A. R. 2009, IEEE Proceedings, 97, 1463,

    2009

  25. [33]

    Zhu, M., McKenna, F., & Scott, M. H. 2018, SoftwareX, 7, 6,

    2018

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.