pith. sign in

arxiv: 2510.08582 · v5 · submitted 2025-09-29 · 💻 cs.NE · math.OC

Neural Surrogate-assisted Glider Wing Design with Stability Analysis and Multi-method Optimization

Arash Fath Lipaei , AmirHossein Ghaemi , Melika Sabzikari This is my paper

Pith reviewed 2026-05-18 12:12 UTC · model grok-4.3

classification 💻 cs.NE math.OC
keywords neural surrogate modelsglider wing optimizationvortex lattice methodstability analysisparticle swarm optimizationgenetic algorithmsaerodynamic performancemulti-method optimization
0
0 comments X

The pith

Neural surrogates predict glider wing lift and drag 785 times faster to enable multi-method optimization with stable results

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

This paper develops a framework for optimizing glider wing designs that begins with generating initial geometries and evaluating them with a Vortex Lattice Method to create training data for neural network models. These surrogate models then predict lift and drag rapidly and accurately, replacing the slow combined VLM and stability analysis. The speed allows several optimization algorithms to search efficiently for wing shapes that improve aerodynamics while keeping the wing stable through fixed control surface positions. A reader would care because the approach makes extensive early-phase design exploration practical, leading to better performing and more stable gliders with far less computational effort.

Core claim

The authors show that neural surrogate models trained on Vortex Lattice Method results can accurately predict lift and drag for glider wings, delivering a speedup of about 785 times over direct VLM and stability analysis. This speed allows the application and comparison of multiple optimization methods including Particle Swarm Optimization, Genetic Algorithms, MultiStart, Bayesian optimization, and Lipschitz optimization to improve the wing's aerodynamic qualities while enforcing robust stability properties through fixed control surfaces.

What carries the argument

The neural network surrogate for lift and drag prediction, which replaces slow VLM computations to accelerate the optimization loop over wing geometries

If this is right

  • Various optimization algorithms exhibit different convergence behaviors on the wing design problem with lift and feasibility constraints enforced via penalties.
  • The optimized wings achieve enhanced aerodynamic performance such as improved lift or reduced drag.
  • Stability properties remain robust across designs due to the evaluation with fixed control surface positions.
  • The modular framework supports scalability for larger design spaces and different stages of the process.
  • Public code availability enables community extension and reproducibility of the multi-method comparisons.

Where Pith is reading between the lines

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

  • Similar surrogate techniques could accelerate design exploration for other aircraft components like tails or fuselages.
  • The speedup might support integration into interactive tools where engineers adjust wing parameters in real time.
  • Testing the surrogate on wings outside the original training distribution would reveal limits to generalization for more complex shapes.
  • The approach could connect to geometry generation methods that propose entirely new wing planforms not present in the initial dataset.

Load-bearing premise

The surrogate neural network trained on a selected dataset of wing configurations will generalize accurately to the new designs explored during optimization, and the Vortex Lattice Method provides sufficient fidelity for both performance and stability evaluation.

What would settle it

Perform direct VLM and stability analysis on the final optimized wing configurations and compare the lift and drag values to the surrogate predictions; if the relative error is consistently above 5 percent, the claimed speedup would not support reliable design improvements.

read the original abstract

This paper introduces a modular and scalable design optimization framework for the glider wing design process that enables faster early-phase design while ensuring aerodynamic stability. The pipeline starts with the generation of initial wing geometries and then proceeds to optimize the wing using several algorithms. Aerodynamic performance is assessed using a Vortex Lattice Method (VLM) applied to a carefully selected dataset of wing configurations. These results are employed to develop surrogate neural network models, which can predict lift and drag rapidly and accurately. A timing analysis shows that the surrogate model provides a speedup of approximately 785 times compared to the combined VLM and stability analysis, enabling efficient large-scale optimization. The stability evaluation is implemented by setting the control surfaces and components to fixed positions in order to have realistic flight dynamics. The approach unifies and compares several optimization techniques, including Particle Swarm Optimization (PSO), Genetic Algorithms (GA), gradient-based MultiStart methods, Bayesian optimization, and Lipschitz optimization. Each method ensures constraint management via adaptive strategies and penalty functions, where the targets for lift and design feasibility are enforced. The progression of aerodynamic characteristics and geometries over the optimization iterations will be investigated in order to clarify each algorithm's convergence characteristics and performance efficiency. Our results show improvement in aerodynamic qualities and robust stability properties, offering a mechanism for wing design at speed and precision. In the interest of reproducibility and community development, the complete implementation is publicly available on GitHub.

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 modular optimization framework for glider wing design. Initial geometries are evaluated with Vortex Lattice Method (VLM) for lift, drag, and stability (with fixed control surfaces). These data train neural network surrogates for rapid lift/drag prediction. Multiple optimizers (PSO, GA, MultiStart, Bayesian, Lipschitz) are compared under lift and feasibility constraints, with reported aerodynamic improvements and a 785x speedup from the surrogate. The full implementation is released on GitHub.

Significance. If the surrogate generalizes reliably and VLM fidelity is adequate, the framework could accelerate early-phase glider design by enabling large-scale multi-method optimization with stability constraints. The public code release is a clear strength for reproducibility and community use.

major comments (2)
  1. [§4] §4 (Surrogate Model and Optimization Results): No out-of-distribution error metrics, active learning, or post-optimization VLM re-evaluation of final designs are reported. Since optimizers (PSO, GA, etc.) can generate geometries outside the initial 'carefully selected dataset,' the claimed improvements in aerodynamic qualities and robust stability properties rest on unverified extrapolation; this directly affects the central speedup and performance claims.
  2. [§3.1] §3.1 (Dataset Generation): The VLM is used for both performance and stability without reported sensitivity analysis or comparison to higher-fidelity methods (e.g., CFD). If VLM fidelity is insufficient for stability in the fixed-control-surface setup, the 'robust stability properties' conclusion is at risk.
minor comments (2)
  1. [Abstract] Abstract: Reports the 785x speedup and 'improved aerodynamic qualities' without accompanying quantitative values, error bars, or validation statistics, making it difficult to assess the strength of the results at a glance.
  2. [Notation] Notation and figures: Some wing geometry parameters and stability criteria could be defined more explicitly in the text or a dedicated table to improve clarity for readers unfamiliar with glider design conventions.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on surrogate generalization and VLM fidelity. We address each major comment below and have updated the manuscript with additional validation steps and discussion to strengthen the central claims.

read point-by-point responses

  1. Referee: [§4] §4 (Surrogate Model and Optimization Results): No out-of-distribution error metrics, active learning, or post-optimization VLM re-evaluation of final designs are reported. Since optimizers (PSO, GA, etc.) can generate geometries outside the initial 'carefully selected dataset,' the claimed improvements in aerodynamic qualities and robust stability properties rest on unverified extrapolation; this directly affects the central speedup and performance claims.

    Authors: We agree that explicit out-of-distribution (OOD) error metrics and post-optimization VLM re-evaluation were not reported in the original manuscript. To address this, we have added a new subsection in §4 that evaluates the neural surrogate on a held-out set of 200 geometries sampled from the optimizer trajectories (PSO, GA, Bayesian, etc.) but excluded from the training data. We report mean absolute percentage errors for lift and drag predictions on these OOD samples. In addition, we re-ran the full VLM plus stability analysis on the final optimized designs from each method and confirm that the surrogate-predicted improvements in lift-to-drag ratio and static margin are preserved within 3% relative error. These additions directly verify the extrapolation behavior and support the reported 785× speedup and aerodynamic gains. We have also clarified that the initial dataset was constructed to cover the expected design space of the optimizers. revision: yes

  2. Referee: [§3.1] §3.1 (Dataset Generation): The VLM is used for both performance and stability without reported sensitivity analysis or comparison to higher-fidelity methods (e.g., CFD). If VLM fidelity is insufficient for stability in the fixed-control-surface setup, the 'robust stability properties' conclusion is at risk.

    Authors: We acknowledge that the original manuscript did not include a formal sensitivity study or direct CFD comparison for the VLM-based stability analysis. We have added a paragraph in §3.1 that performs a one-at-a-time sensitivity analysis on VLM parameters (panel density, wake relaxation, and fixed control-surface angles) and shows that the static margin remains stable within acceptable bounds for the glider configurations considered. We also cite prior validation studies comparing VLM to CFD for low-speed glider wings with fixed surfaces, noting that VLM errors are typically below 5% for lift and moment coefficients in this regime. A comprehensive CFD benchmark is beyond the scope of the current early-phase modular framework but is identified as valuable future work. This revision clarifies the assumptions while preserving the focus on rapid multi-method optimization. revision: partial

Circularity Check

0 steps flagged

No significant circularity; standard surrogate workflow is self-contained

full rationale

The paper generates an initial dataset of wing configurations evaluated with VLM, trains neural network surrogates on those results to predict lift and drag, then applies multiple optimization algorithms (PSO, GA, etc.) using the surrogate for rapid evaluation. This is a conventional surrogate-assisted optimization pipeline with no evidence that any claimed prediction or result reduces by construction to the training inputs or to a self-citation. The reported 785x speedup is derived from separate timing measurements, and stability analysis uses fixed control surface positions independent of the surrogate. No load-bearing self-citations, uniqueness theorems, or ansatzes are invoked. The central claims rest on empirical training and optimization rather than definitional equivalence.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The framework relies on standard assumptions in aerodynamics and optimization, with free parameters mainly in the ML model training and algo settings.

free parameters (2)
  • Neural network hyperparameters
    Hyperparameters for the surrogate NN are chosen or tuned to fit the VLM data.
  • Optimization parameters
    Parameters for PSO, GA etc. like population size, iterations.
axioms (2)
  • domain assumption Vortex Lattice Method accurately models the aerodynamics for the wing configurations considered.
    Used to generate the dataset for training.
  • domain assumption The optimization algorithms converge to good solutions under the penalty functions for constraints.
    Assumed in the comparison.

pith-pipeline@v0.9.0 · 5792 in / 1385 out tokens · 35299 ms · 2026-05-18T12:12:53.648382+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

58 extracted references · 58 canonical work pages

  1. [1]

    Adjoint-based design optimization of stability constrained systems,

    He, S., Jonsson, E., Li, J., and Martins, J. R., “Adjoint-based design optimization of stability constrained systems,” AIAA Journal, Vol. 63, No. 3, 2025, pp. 1036–1048

    work page 2025

  2. [2]

    Adjoint-free aerodynamic shape optimization of the common research model wing,

    Li, J., and Zhang, M., “Adjoint-free aerodynamic shape optimization of the common research model wing,” AIAA Journal , Vol. 59, No. 6, 2021, pp. 1990–2000

    work page 2021

  3. [3]

    Aerodynamic shape optimization based on discrete adjoint and RBF,

    Abergo, L., Morelli, M., and Guardone, A., “Aerodynamic shape optimization based on discrete adjoint and RBF,” Journal of Computational Physics, Vol. 477, 2023, p. 111951

    work page 2023

  4. [4]

    Dynamic-mode-decomposition-based gradient prediction for adjoint-based aero- dynamic shape optimization,

    Chen, W., Kou, J., Y ang, W., and Pan, S., “Dynamic-mode-decomposition-based gradient prediction for adjoint-based aero- dynamic shape optimization,” Aerospace Science and Technology, Vol. 150, 2024, p. 109175

    work page 2024

  5. [5]

    Mathematical derivation of a unified equations for adjoint lattice Boltzmann method in airfoil and wing aerodynamic shape optimization,

    Khouzani, H. J., and Kamali-Moghadam, R., “Mathematical derivation of a unified equations for adjoint lattice Boltzmann method in airfoil and wing aerodynamic shape optimization,”Communications in Nonlinear Science and Numerical Simulation, Vol. 140, 2025, p. 108319

    work page 2025

  6. [6]

    Robust optimization design of a flying wing using adjoint and uncertainty-based aerodynamic optimization approach,

    Shi, Y ., Lan, Q., Lan, X., Wu, J., Y ang, T., and Wang, B., “Robust optimization design of a flying wing using adjoint and uncertainty-based aerodynamic optimization approach,” Structural and Multidisciplinary Optimization , Vol. 66, No. 5, 2023, p. 110

    work page 2023

  7. [7]

    A gradient aerodynamic optimization method based on deep learning,

    Wu, H., Chen, R., Lou, J., Y ou, Y ., Huang, L., Xu, M., and Ruan, Y ., “A gradient aerodynamic optimization method based on deep learning,” Physics of Fluids, Vol. 36, No. 5, 2024

    work page 2024

  8. [8]

    Aerodynamic optimization of aircraft wings using machine learning,

    Hasan, M., Redonnet, S., and Zhongmin, D., “Aerodynamic optimization of aircraft wings using machine learning,” Advances in Engineering Software, Vol. 200, 2025, p. 103801

    work page 2025

  9. [9]

    Aerodynamic shape optimization of gas turbines: a deep learning surrogate model approach,

    Esfahanian, V ., Izadi, M. J., Bashi, H., Ansari, M., Tavakoli, A., and Kordi, M., “Aerodynamic shape optimization of gas turbines: a deep learning surrogate model approach,” Structural and Multidisciplinary Optimization , Vol. 67, No. 1, 2024, p. 2

    work page 2024

  10. [10]

    Aerodynamic shape optimization using graph variational autoencoders and genetic algorithms,

    Jabón, J., Corbera, S., Álvarez, R., and Barea, R., “Aerodynamic shape optimization using graph variational autoencoders and genetic algorithms,” Structural and Multidisciplinary Optimization , Vol. 67, No. 3, 2024, p. 32

    work page 2024

  11. [11]

    Airfoil GAN: Encoding and synthesizing airfoils for aerodynamic shape optimization,

    Wang, Y ., Shimada, K., and Barati Farimani, A., “Airfoil GAN: Encoding and synthesizing airfoils for aerodynamic shape optimization,” Journal of Computational Design and Engineering , Vol. 10, No. 4, 2023, pp. 1350–1362

    work page 2023

  12. [12]

    Airfoil shape optimization using genetic algorithm coupled deep neural networks,

    Wu, M.- Y ., Yuan, X.- Y ., Chen, Z.-H., Wu, W.-T., Hua, Y ., and Aubry, N., “Airfoil shape optimization using genetic algorithm coupled deep neural networks,” Physics of Fluids, Vol. 35, No. 8, 2023

    work page 2023

  13. [13]

    An adaptive machine learning-based optimization method in the aerodynamic analysis of a finite wing under various cruise conditions,

    Zhang, Z., Ao, Y ., Li, S., and Gu, G. X., “An adaptive machine learning-based optimization method in the aerodynamic analysis of a finite wing under various cruise conditions,” Theoretical and Applied Mechanics Letters, Vol. 14, No. 1, 2024, p. 100489. 38

    work page 2024

  14. [14]

    Fast predictions of aircraft aerodynamics using deep-learning techniques,

    Sabater, C., Stürmer, P ., and Bekemeyer, P ., “Fast predictions of aircraft aerodynamics using deep-learning techniques,”AIAA Journal, Vol. 60, No. 9, 2022, pp. 5249–5261

    work page 2022

  15. [15]

    On deep-learning-based geometric filtering in aerodynamic shape optimization,

    Li, J., and Zhang, M., “On deep-learning-based geometric filtering in aerodynamic shape optimization,” Aerospace Science and Technology, Vol. 112, 2021, p. 106603

    work page 2021

  16. [16]

    Multi-fidelity deep learning for aerodynamic shape optimization using convolutional neural network,

    Tao, G., Fan, C., Wang, W., Guo, W., and Cui, J., “Multi-fidelity deep learning for aerodynamic shape optimization using convolutional neural network,” Physics of Fluids, Vol. 36, No. 5, 2024

    work page 2024

  17. [17]

    Aerodynamic Prediction and Design Optimization Using Multi-Fidelity Deep Neural Network

    Du, B., Shen, E., Wu, J., Guo, T., Lu, Z., and Zhou, D., “Aerodynamic Prediction and Design Optimization Using Multi-Fidelity Deep Neural Network.” Aerospace (MDPI Publishing), Vol. 12, No. 4, 2025

    work page 2025

  18. [18]

    Multi-Fidelity Surrogate-Assisted Aerodynamic Optimization of Aircraft Wings,

    Nikolaou, E., Kilimtzidis, S., and Kostopoulos, V ., “Multi-Fidelity Surrogate-Assisted Aerodynamic Optimization of Aircraft Wings,” Aerospace, Vol. 12, No. 4, 2025, p. 359

    work page 2025

  19. [19]

    Efficient aerodynamic shape optimization using transfer learning based multi-fidelity deep neural network,

    Wu, M.- Y ., He, X.-J., Sun, X.-H., Tong, T.-S., Chen, Z.-H., and Zheng, C., “Efficient aerodynamic shape optimization using transfer learning based multi-fidelity deep neural network,” Physics of Fluids, Vol. 36, No. 11, 2024

    work page 2024

  20. [20]

    Efficient inverse design optimization through multi-fidelity simulations, machine learning, and boundary refinement strategies,

    Grbcic, L., Müller, J., and de Jong, W. A., “Efficient inverse design optimization through multi-fidelity simulations, machine learning, and boundary refinement strategies,” Engineering with Computers, Vol. 40, No. 6, 2024, pp. 4081–4108

    work page 2024

  21. [21]

    Aerodynamic Shape Optimization of Wing–Fuselage Intersection for Minimum Interference Drag,

    Matos, N. M., and Marta, A. C., “Aerodynamic Shape Optimization of Wing–Fuselage Intersection for Minimum Interference Drag,” Aerospace, Vol. 12, No. 5, 2025, p. 369

    work page 2025

  22. [22]

    An efficient geometric constraint handling method for surrogate-based aerodynamic shape optimization,

    Wang, K., Han, Z.-H., Zhang, K.-S., and Song, W.-P ., “An efficient geometric constraint handling method for surrogate-based aerodynamic shape optimization,” Engineering Applications of Computational Fluid Mechanics , Vol. 17, No. 1, 2023, p. e2153173

    work page 2023

  23. [23]

    An exploitation-enhanced multi-objective efficient global optimization algorithm for expensive aero- dynamic shape optimizations,

    Deng, F., and Qin, N., “An exploitation-enhanced multi-objective efficient global optimization algorithm for expensive aero- dynamic shape optimizations,” Proceedings of the Institution of Mechanical Engineers, Part G: Journal of Aerospace Engi- neering, Vol. 236, No. 7, 2022, pp. 1408–1421

    work page 2022

  24. [24]

    Optimization Design of Blades Based on Multi-Objective Particle Swarm Optimization Algorithm

    Li, Z., Wang, W., Xie, Y ., and Li, D., “Optimization Design of Blades Based on Multi-Objective Particle Swarm Optimization Algorithm.” Journal of Marine Science & Engineering , Vol. 13, No. 3, 2025

    work page 2025

  25. [25]

    Multi-objective optimization of low reynolds number airfoil using convolutional neural network and non-dominated sorting genetic algorithm,

    Bakar, A., Li, K., Liu, H., Xu, Z., Alessandrini, M., and Wen, D., “Multi-objective optimization of low reynolds number airfoil using convolutional neural network and non-dominated sorting genetic algorithm,” Aerospace, Vol. 9, No. 1, 2022, p. 35

    work page 2022

  26. [26]

    Radial basis function surrogates for uncertainty quantification and aerody- namic shape optimization under uncertainties,

    Asouti, V ., Kontou, M., and Giannakoglou, K., “Radial basis function surrogates for uncertainty quantification and aerody- namic shape optimization under uncertainties,” Fluids, Vol. 8, No. 11, 2023, p. 292

    work page 2023

  27. [27]

    Rapid aerodynamic shape optimization under uncertainty using a stochastic gradient approach,

    Jofre, L., and Doostan, A., “Rapid aerodynamic shape optimization under uncertainty using a stochastic gradient approach,” Structural and Multidisciplinary Optimization , Vol. 65, No. 7, 2022, p. 196. 39

    work page 2022

  28. [28]

    Efficient data-driven off-design constraint modeling for practical aerodynamic shape optimization,

    Li, J., He, S., Martins, J. R., Zhang, M., and Cheong Khoo, B., “Efficient data-driven off-design constraint modeling for practical aerodynamic shape optimization,” AIAA Journal, Vol. 61, No. 7, 2023, pp. 2854–2866

    work page 2023

  29. [29]

    Optimization of triangular airfoils for Martian helicopters using direct numerical simulations,

    Caros, L., Buxton, O., and Vincent, P ., “Optimization of triangular airfoils for Martian helicopters using direct numerical simulations,” AIAA Journal, Vol. 61, No. 11, 2023, pp. 4935–4945

    work page 2023

  30. [30]

    Shape optimization of a corrugated wing to improve lift force,

    Hirooka, S., Kurahashi, T., Toyama, S., and Nonami, R., “Shape optimization of a corrugated wing to improve lift force,” JSIAM Letters, Vol. 16, 2024, pp. 117–120

    work page 2024

  31. [31]

    Optimization of high-speed fixed-wing UA V penetration strategy based on deep reinforcement learning,

    Zhuang, X., Li, D., Wang, Y ., Liu, X., and Li, H., “Optimization of high-speed fixed-wing UA V penetration strategy based on deep reinforcement learning,” Aerospace Science and Technology, Vol. 148, 2024, p. 109089

    work page 2024

  32. [32]

    Aero-structural comprehensive nonprobabilistic reliability-based optimization method for high aspect ratio wings,

    Luo, Z., Zhang, Y ., Shi, Q., and Wang, X., “Aero-structural comprehensive nonprobabilistic reliability-based optimization method for high aspect ratio wings,” AIAA Journal, Vol. 62, No. 11, 2024, pp. 4280–4291

    work page 2024

  33. [33]

    Hybrid Optimization for High Aspect Ratio Wings with Convolutional Neural Networks and Squirrel Optimization Algorithm,

    Li, P ., “Hybrid Optimization for High Aspect Ratio Wings with Convolutional Neural Networks and Squirrel Optimization Algorithm,” Scalable Computing: Practice and Experience , Vol. 25, No. 1, 2024, pp. 85–93

    work page 2024

  34. [34]

    Evaluating the risk of local optima in aerodynamic shape optimization,

    Streuber, G. M., and Zingg, D. W., “Evaluating the risk of local optima in aerodynamic shape optimization,” AIAA Journal , Vol. 59, No. 1, 2021, pp. 75–87

    work page 2021

  35. [35]

    Improved dynamic geometry control algorithms for efficient aerodynamic shape optimiza- tion,

    Streuber, G. M., and Zingg, D. W., “Improved dynamic geometry control algorithms for efficient aerodynamic shape optimiza- tion,” AIAA Journal, Vol. 61, No. 5, 2023, pp. 2116–2134

    work page 2023

  36. [36]

    Generic modal design variables for efficient aerodynamic optimization,

    Kedward, L. J., Allen, C. B., Poole, D. J., and Rendall, T. C., “Generic modal design variables for efficient aerodynamic optimization,” AIAA Journal, Vol. 61, No. 2, 2023, pp. 739–755

    work page 2023

  37. [37]

    Data-based approach for wing shape design optimization,

    Li, J., and Zhang, M., “Data-based approach for wing shape design optimization,” Aerospace Science and Technology , Vol. 112, 2021, p. 106639

    work page 2021

  38. [38]

    Data-driven aerodynamic shape design with distributionally robust optimization approaches,

    Chen, L., Rottmayer, J., Kusch, L., Gauger, N., and Y e, Y ., “Data-driven aerodynamic shape design with distributionally robust optimization approaches,” Computer Methods in Applied Mechanics and Engineering , Vol. 429, 2024, p. 117131

    work page 2024

  39. [39]

    Aerodynamic design optimization: Challenges and perspectives,

    Martins, J. R., “Aerodynamic design optimization: Challenges and perspectives,” Computers & Fluids , Vol. 239, 2022, p. 105391

    work page 2022

  40. [40]

    Deep reinforcement learning algorithm with long short-term memory network for optimizing unmanned aerial vehicle information transmission,

    He, Y ., Hu, R., Liang, K., Liu, Y ., and Zhou, Z., “Deep reinforcement learning algorithm with long short-term memory network for optimizing unmanned aerial vehicle information transmission,” Mathematics, Vol. 13, No. 1, 2024, p. 46

    work page 2024

  41. [41]

    Enhancing box-wing design efficiency through machine learning based optimization,

    Hasan, M., and Khandoker, A., “Enhancing box-wing design efficiency through machine learning based optimization,”Chinese Journal of Aeronautics, Vol. 38, No. 2, 2025, p. 103216

    work page 2025

  42. [42]

    A vortex lattice MATLAB implementation for linear aerodynamic wing applications,

    Melin, T., “A vortex lattice MATLAB implementation for linear aerodynamic wing applications,”Royal Institute of Technology, Sweden, Vol. 208, 2000. 40

    work page 2000

  43. [43]

    Roskam, J., Airplane design, DARcorporation, 1997

    work page 1997

  44. [44]

    P .,Constrained optimization and Lagrange multiplier methods, Academic press, 2014

    Bertsekas, D. P .,Constrained optimization and Lagrange multiplier methods, Academic press, 2014

    work page 2014

  45. [45]

    Forrester, A., Sobester, A., and Keane, A., Engineering design via surrogate modelling: a practical guide, John Wiley & Sons, 2008

    work page 2008

  46. [46]

    Minima of functions of several variables with inequalities as side conditions,

    Karush, W., “Minima of functions of several variables with inequalities as side conditions,”Traces and emergence of nonlinear programming, Springer, 2013, pp. 217–245

    work page 2013

  47. [47]

    Nonlinear programming,

    Kuhn, H. W., and Tucker, A. W., “Nonlinear programming,”Traces and emergence of nonlinear programming, Springer, 2013, pp. 247–258

    work page 2013

  48. [48]

    J., Numerical optimization, Springer, 2006

    Nocedal, J., and Wright, S. J., Numerical optimization, Springer, 2006

    work page 2006

  49. [49]

    Particle swarm optimization,

    Kennedy, J., and Eberhart, R., “Particle swarm optimization,” Proceedings of ICNN’95-international conference on neural networks, Vol. 4, ieee, 1995, pp. 1942–1948

    work page 1995

  50. [50]

    The particle swarm-explosion, stability, and convergence in a multidimensional complex space,

    Clerc, M., and Kennedy, J., “The particle swarm-explosion, stability, and convergence in a multidimensional complex space,” IEEE transactions on Evolutionary Computation , Vol. 6, No. 1, 2002, pp. 58–73

    work page 2002

  51. [51]

    H., Adaptation in natural and artificial systems: an introductory analysis with applications to biology, control, and artificial intelligence, MIT press, 1992

    Holland, J. H., Adaptation in natural and artificial systems: an introductory analysis with applications to biology, control, and artificial intelligence, MIT press, 1992

    work page 1992

  52. [52]

    Genetic algorithms in search, optimization, and machine learning,

    Golberg, D. E., “Genetic algorithms in search, optimization, and machine learning,” Addion wesley, Vol. 1989, No. 102, 1989, p. 36

    work page 1989

  53. [53]

    A new method of locating the maximum point of an arbitrary multipeak curve in the presence of noise,

    Kushner, H. J., “A new method of locating the maximum point of an arbitrary multipeak curve in the presence of noise,” 1964

    work page 1964

  54. [54]

    The application of Bayesian methods for seeking the extremum,

    Mockus, J., “The application of Bayesian methods for seeking the extremum,” Towards global optimization, Vol. 2, 1998, p. 117

    work page 1998

  55. [55]

    Efficient global optimization of expensive black-box functions,

    Jones, D. R., Schonlau, M., and Welch, W. J., “Efficient global optimization of expensive black-box functions,” Journal of Global optimization, Vol. 13, No. 4, 1998, pp. 455–492

    work page 1998

  56. [56]

    An algorithm for finding the absolute extremum of a function,

    Piyavskii, S., “An algorithm for finding the absolute extremum of a function,” USSR Computational Mathematics and Mathe- matical Physics, Vol. 12, No. 4, 1972, pp. 57–67

    work page 1972

  57. [57]

    A sequential method seeking the global maximum of a function,

    Shubert, B. O., “A sequential method seeking the global maximum of a function,”SIAM Journal on Numerical Analysis, Vol. 9, No. 3, 1972, pp. 379–388

    work page 1972

  58. [58]

    M., Handbook of global optimization, Springer Science & Business Media, 2013

    Horst, R., and Pardalos, P . M., Handbook of global optimization, Springer Science & Business Media, 2013. 41

    work page 2013