Comparative analysis of modern approaches to optical system design automation

Chertov A.N., Khokhlov D.D. Comparative analysis of modern approaches to optical system design automation. Scientific and Technical Journal of Information Technologies, Mechanics and Optics, 2026, vol. 26, no. 1, pp. 35–41 (in Russian). doi: 10.17586/2226-1494-2026-26-1-35-41

Automation of optical system design is one of the key directions in modern optical engineering. The combination of physically grounded simulators, numerical optimization techniques, and machine-learning algorithms enables the development of compact, energy-efficient, and manufacturable optical systems. However, the high dimensionality of parameter spaces, significant computational costs, and the lack of unified verification criteria necessitate a systematic analysis of the applicability of different methods. The article presents a systematization of current approaches to optical system design automation, an analysis of their key characteristics, and an assessment of prospects for further development. Five main classes of methods are examined: differentiable physical models, deep-learning algorithms, evolutionary and metaheuristic optimizers, hybrid schemes combining machine learning methods and physics-based modeling, and fully forward (optical) training approaches. Consideration is given to the formation of a unified comparison framework that enables objective evaluation of speed, accuracy, reliability, robustness, generalization capability, computational complexity, and energy efficiency across different algorithms. A classification of automated optical design methods is proposed. The analysis includes physical models, neural-network architectures, and optimization algorithms. A comparative evaluation based on a unified set of metrics is provided, including qualitative and quantitative assessments derived from peer-reviewed publications from 2019 to 2025. The study demonstrates that differentiable physical methods provide the highest level of physical fidelity and accuracy. Deep-learning methods ensure maximal speed of solution generation. Evolutionary algorithms exhibit robustness against local minima. Hybrid approaches offer an effective balance between speed and physical correctness. Fully forward training methods (FFM approaches) and optical neural networks deliver high energy efficiency and show potential for hardware acceleration of the design process. The results can serve as a basis for selecting design strategies for optical systems across tasks of varying complexity — from preliminary configuration search to high-precision optimization. Progress in automated optical system design is linked to further standardization of datasets, integration of hybrid approaches, and the advancement of optical neural networks.

1. Wu C.-M., Tang X.-X., Xia Y.-Y., Yang H.-X., Xu F.-J. High precision ray tracing method for space camera in optical design. Acta Physica Sinica, 2023, vol. 72,no. 8,pp. 084201. (in Chinese). https://doi.org/10.7498/aps.72.20222463

2. Liu X.B., Zhang X.X., Fu T.J., Wang K.Z., Sun F.K., Bai T.Z., Wang D.Global information selectively guided gradient descent for ab initio optical design. Optics and Laser Technology,2025,vol. 184,pp. 112497. https://doi.org/10.1016/j.optlastec.2025.112497

3. Zhang W., Ren Z., Zhou J., Chen S., Feng H., Li Q., Xu Z., Chen Y. End-to-end automatic lens design with a differentiable diffraction model. Optics Express, 2024, vol. 32, no. 25, pp. 44328–44345. https://doi.org/10.1364/oe.540590

4. Ivanova T.V., Romanova G.E., Zhukova T.I., Kalinkina O.S. Automation of athermal cemented doublet synthesis. Scientific and Technical Journal of Information Technologies, Mechanics and Optics, 2019, vol. 19, no. 4, pp. 594–601. (in Russian). https://doi.org/10.17586/2226-1494-2019-19-4-594-601

5. Nguyen DH, Bakholdin AB. Automation of synthesis and ranking of cemented and air-spaced doublets. Computer Optics, 2022, vol. 46, no. 1, pp. 83-89. (in Russian). https://doi.org/10.18287/2412-6179-CO-923

6. Nosov P.A., Pavlov V.Yu., Pakhomov I.I., Shirankov A.F.Aberrational synthesis of optical systems intended for the conversion of laser beams. Journal of Optical Technology, 2011, vol. 78, no. 9, pp. 586–593. https://doi.org/10.1364/JOT.78.000586

7. Livshits I.L., Tochilina T.V., Faehnle O., Volkova S.L. Design strategy and management of aberration correction process for lens with high complexity index. Scientific and Technical Journal of Information Technologies, Mechanics and Optics,2021, vol. 21, no. 1, pp. 40–51. (in Russian). https://doi.org/10.17586/2226-1494-2021-21-1-40-51

8. Côté G., Lalonde J.-F., Thibault S. Deep learning-enabled framework for automatic lens design starting point generation. Optics Express, 2021, vol. 29, no. 3, pp. 3841–3854. https://doi.org/10.1364/oe.401590

9. Luo M., Lee S.-S. Inverse design of optical lenses enabled by generative flow-based invertible neural networks. Scientific Reports, 2023,vol. 13, no. 1, pp. 16416. https://doi.org/10.1038/s41598-023-43698-3

10. Fu T., Zhang J., Sun R., Huang Y., Xu W., Yang S., Zhu Z., Chen H. Optical neural networks: progress and challenges. Light Science and Applications, 2024, vol. 13, no. 1, pp. 263. https://doi.org/10.1038/s41377-024-01590-3

11. Höschel K., Lakshminarayanan V. Genetic algorithms for lens design: a review. Journal of Optics, 2019, vol. 48, no. 1, pp. 134–144. https://doi.org/10.1007/s12596-018-0497-3

12. Yang X., Fu Q., Heidrich W. Curriculum learning for ab initio deep learned refractive optics. Nature Communications, 2024, vol. 15, no. 1, pp. 6572. https://doi.org/10.1038/s41467-024-50835-7

13. Zhao J., Liang K., Van Vleck A., Du L., Winetraub Y., Peng Y.F., de la Zerda A. Extending depth-of-field of arbitrary diffractive optics with needle-shaped beam modulation. ACS Photonics, 2024, vol. 11,no. 9, pp. 3481–3490. https://doi.org/10.1021/acsphotonics.4c00260

14. Yow A.P., Wong D., Zhang Y., Menke C., Wolleschensky R., Török P. Artificial intelligence in optical lens design. Artificial Intelligence Review, 2024, vol. 57, no. 8, pp. 193. https://doi.org/10.1007/s10462-024-10842-y

15. Nesterov G.V., Khokhlov D.D., Batshev V.I. Automation of cemented doublet design using machine learning methods. Physical Bases of Instrumentation, 2024, vol. 13, no.  4 (54). pp. 57–63. (in Russian). https://doi.org/10.25210/jfop-2404-INGPBM

16. Volatier J.-B., Beaussier S.J., Druart G., Jougla P., Keller F. Implementation of FORMIDABLE: a generalized differential optical design library with NURBS capabilities. Journal of the European Optical Society-Rapid Publications, 2024, vol. 20, no. 1, pp. 2. https://doi.org/10.1051/jeos/2023043

17. Xue Z., Zhou T., Xu Z., Yu S., Dai Q., Fang L. Fully forward mode training for optical neural networks. Nature, 2024, vol. 632, no. 8024, pp. 280–286. https://doi.org/10.1038/s41586-024-07687-4

18. Wang Z., Peng Y., Fang L., Gao L. Computational optical imaging: on the convergence of physical and digital layers. Optica, 2025, vol. 12, no. 1, pp. 113–130. https://doi.org/10.1364/OPTICA.544943

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License