Artificial Intelligence in Generative Design: A Structured Review of Trends and Opportunities in Techniques and Applications
Abstract
1. Introduction
2. Related Work
3. Review Methodology
3.1. Initial Paper Gathering
- Methodological research aiming to synthesise designs;
- An application of synthesis of “function and form” engineering designs;
- Use of AI techniques to directly inform or perform design synthesis.
3.2. Down-Selection
3.3. Coding
- What are the key trends, drivers, and application areas at the intersection of AI and Generative Design?
- What issues, limitations, and barriers are reported in the existing literature?
- What gaps, emerging opportunities, and future challenges exist for researchers and practitioners?
3.4. Analysis
4. Results
4.1. Initial Corpus
4.2. Down-Selection
4.3. Trends and Code Analysis
4.3.1. Bibliometric Trend Analysis
4.3.2. Techniques
4.3.3. AI Function
4.3.4. Applications
4.3.5. Benefit Themes
4.3.6. Limitation Themes
4.3.7. Future Work
5. Discussion
5.1. Technique Trends
5.2. Application Trends
5.3. Technique–Function Pairings
5.4. Opportunities
5.5. Evaluation of Our Approach
6. Conclusions
- A comprehensive survey of trends in the use of AI and GD techniques in a wide range of applications. The key findings from this are:
- There has been a significant increase in research activity at this intersection, with 72.7% relevant articles published since 2021 compared to the period from 2008 to 2020. This reflects both growing interest in the field and technological advancement.
- The analysis reveals two primary applications of AI in GD: as surrogate models to accelerate existing optimisation processes and for direct design generation.
- Although MLPs remains the most commonly used AI technique (34% of the articles), there has been a shift toward modern approaches like GANs (23.4%) and CNNs (19.1%) in the past 5 years. MLPs in recent research tended to be used as components in larger systems. This evolution mirrors broader trends in AI research, although with a lag between advances in general AI and their application to engineering design.
- The review identified persistent challenges across the field. Data generation and training costs were cited as limitations in 56.8% of the articles, while model accuracy concerns were raised in 30%. These challenges particularly affect surrogate modelling approaches, where researchers must carefully balance the computational costs of training against potential performance benefits.
- Identification of opportunities and priorities for further research on the use of AI to enhance GD:
- Training of foundation surrogate models capable of generalising to multiple scenarios could address limitations in computational cost and accuracy.
- The development of GenAI systems capable of synthesis using structured engineering data formats.
- Investigation into the use of unsupervised learning systems to better capture and reproduce the physical meaning behind design features.
- Improvement in the manufacturability of GD outcomes.
Author Contributions
Funding
Data Availability Statement
Conflicts of Interest
References
- Schmit, L.A. Structural Design by Systematic Synthesis. In Proceedings of the Second National Conference on Electronic Computation; ASCE: Reston, VA, USA, 1960. [Google Scholar]
- Caetano, I.; Santos, L.; Leitão, A. Computational Design in Architecture: Defining Parametric, Generative, and Algorithmic Design. Front. Archit. Res. 2020, 9, 287–300. [Google Scholar] [CrossRef]
- Huang, M.; Cui, T.; Liu, C.; Du, Z.; Zhang, J.; He, C.; Guo, X. A Problem-Independent Machine Learning (PIML) Enhanced Substructure-Based Approach for Large-Scale Structural Analysis and Topology Optimization of Linear Elastic Structures. Extrem. Mech. Lett. 2023, 63, 102041. [Google Scholar] [CrossRef]
- Wang, Z.; Melkote, S.; Rosen, D.W. Generative Design by Embedding Topology Optimization into Conditional Generative Adversarial Network. J. Mech. Des. 2023, 145, 111702. [Google Scholar] [CrossRef]
- Parekh, V.; Flore, D.; Schops, S. Variational Autoencoder-Based Metamodeling for Multi-Objective Topology Optimization of Electrical Machines. IEEE Trans. Magn. 2022, 58, 7500604. [Google Scholar] [CrossRef]
- Shrestha, P.R.; Timalsina, D.; Bista, S.; Shrestha, B.P.; Shakya, T.M. Generative Design Approach for Product Development. In Proceedings of the the 7th International Conference on Engineering, Applied Sciences and Technology: (ICEAST2021), Bangkok, Thailand, 1–3 April 2021; p. 020008. [Google Scholar] [CrossRef]
- Mihály, C.; Ildikó, D.M. Mass Reduction of the Upright of a Racing Car with Innovative Methods. Acta Polytech. Hung. 2024, 21, 253–264. [Google Scholar] [CrossRef]
- Ntintakis, I.; Stavroulakis, G.E.; Sfakianakis, G.; Fiotodimitrakis, N. Utilizing Generative Design for Additive Manufacturing. In Recent Advances in Manufacturing Processes and Systems; Dave, H.K., Dixit, U.S., Nedelcu, D., Eds.; Springer Nature: Singapore, 2022; pp. 977–989. [Google Scholar] [CrossRef]
- Peckham, O.; Elverum, C.W.; Hicks, B.; Goudswaard, M.; Snider, C.; Steinert, M.; Eikevåg, S.W. Investigating and Characterizing the Systemic Variability When Using Generative Design for Additive Manufacturing. Appl. Sci. 2024, 14, 4750. [Google Scholar] [CrossRef]
- Research, S. Global Generative Design Market Size, Trends, Forecast to 2030. Available online: https://straitsresearch.com/report/generative-design-market (accessed on 12 November 2024).
- Elsevier. Scopus—Document Search. Available online: https://www.scopus.com/search/form.uri?display=basic&zone=header&origin=#basic (accessed on 1 November 2024).
- Yüksel, N.; Börklü, H.R.; Sezer, H.K.; Canyurt, O.E. Review of Artificial Intelligence Applications in Engineering Design Perspective. Eng. Appl. Artif. Intell. 2023, 118, 105697. [Google Scholar] [CrossRef]
- Krizhevsky, A.; Sutskever, I.; Hinton, G.E. ImageNet Classification with Deep Convolutional Neural Networks. Commun. ACM 2017, 60, 84–90. [Google Scholar] [CrossRef]
- He, Z.; Wang, Y.H.; Zhang, J. Generative AIBIM: An Automatic and Intelligent Structural Design Pipeline Integrating BIM and Generative AI. Inf. Fusion 2025, 114, 102654. [Google Scholar] [CrossRef]
- Sengar, S.S.; Hasan, A.B.; Kumar, S.; Carroll, F. Generative Artificial Intelligence: A Systematic Review and Applications. Multimed. Tools Appl. 2024. [Google Scholar] [CrossRef]
- BendsOe, M.P. Optimal Shape Design as a Material Distribution Problem. Struct. Optim. 1989, 1, 193–202. [Google Scholar] [CrossRef]
- Bernal, M.; Haymaker, J.R.; Eastman, C. On the Role of Computational Support for Designers in Action. Des. Stud. 2015, 41, 163–182. [Google Scholar] [CrossRef]
- Jenis, J.; Ondriga, J.; Hrcek, S.; Brumercik, F.; Cuchor, M.; Sadovsky, E. Engineering Applications of Artificial Intelligence in Mechanical Design and Optimization. Machines 2023, 11, 577. [Google Scholar] [CrossRef]
- Regenwetter, L.; Nobari, A.H.; Ahmed, F. Deep Generative Models in Engineering Design: A Review. J. Mech. Des. 2022, 144, 071704. [Google Scholar] [CrossRef]
- Khanolkar, P.M.; Vrolijk, A.; Olechowski, A. A Case Study of the Decision Making Behind the Automation of a Composites Based Design Process. Proc. Des. Soc. 2023, 3, 49–58. [Google Scholar] [CrossRef]
- Kretzschmar, M.; Dammann, M.P.; Schwoch, S.; Braun, F.; Saske, B.; Paetzold-Byhain, K. Evaluating the Role of Generative AI in Product Development and Design—A Systematic Review. In Proceedings of the DS 130: Proceedings of NordDesign 2024, Reykjavik, Iceland, 12–14 August 2024; pp. 21–30. [Google Scholar] [CrossRef]
- Shin, S.; Shin, D.; Kang, N. Topology Optimization via Machine Learning and Deep Learning: A Review. J. Comput. Des. Eng. 2023, 10, 1736–1766. [Google Scholar] [CrossRef]
- Cheng, H.; Yi, W.; Ji, L. Multi-Point Optimization Design of High Pressure-Ratio Centrifugal Impeller Based on Machine Learning. In Proceedings of the Volume 2D: Turbomachinery, Virtual, Online, 21–25 September 2020; p. V02DT38A009. [Google Scholar] [CrossRef]
- Singh, A.; D’Arcy, M.; Cohan, A.; Downey, D.; Feldman, S. SciRepEval: A Multi-Format Benchmark for Scientific Document Representations. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, Singapore, 6–10 December 2023; pp. 5548–5566. [Google Scholar] [CrossRef]
- Asmussen, C.B.; Møller, C. Smart Literature Review: A Practical Topic Modelling Approach to Exploratory Literature Review. J. Big Data 2019, 6, 93. [Google Scholar] [CrossRef]
- Aguirre, A.M.G.; Robledo, S.; Zuluaga, M. Topic Modeling: Perspectives From a Literature Review. IEEE Access 2023, 11, 4066–4078. [Google Scholar] [CrossRef]
- Wong, A.; Cheng, F.W.Y.; Keung, A.; Hercules, Y.; Garcia, M.A.; Lim, Y.W.; Pham, L. An Iterative Approach to Topic Modelling. arXiv 2024, arXiv:2407.17892. [Google Scholar] [CrossRef]
- Li, D.; Wu, K.; Lei, V.L. Applying Topic Modeling to Literary Analysis: A Review. Digit. Stud. Lang. Lit. 2024, 1, 113–141. [Google Scholar] [CrossRef]
- Lloyd, S. Least Squares Quantization in PCM. IEEE Trans. Inf. Theory 1982, 28, 129–137. [Google Scholar] [CrossRef]
- MacQueen, J. Some Methods for Classification and Analysis of Multivariate Observations. In Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, Volume 1: Statistics, Berkeley, CA, USA, 21 June–18 July 1965; University of California Press: Berkeley, CA, USA, 1967; Volume 5, pp. 281–298. [Google Scholar]
- Arvai, K. Zenodo Kneed. 2020. Available online: https://doi.org/10.5281/zenodo.6496267 (accessed on 14 January 2025).
- Xu, J.; Wang, L.; Yuan, J.; Shi, J.; Wang, Z.; Zhang, B.; Luo, Z.; Tan, A.C. A Cost-Effective CNN-BEM Coupling Framework for Design Optimization of Horizontal Axis Tidal Turbine Blades. Energy 2023, 282, 128707. [Google Scholar] [CrossRef]
- Li, Y.; Yi, Y.K. Optimal Shape Design Using Machine Learning for Wind Energy and Pressure. J. Build. Eng. 2023, 70, 106337. [Google Scholar] [CrossRef]
- Gisbert, F.; Cadrecha, D.; Quintanal, J.; Sotillo, A.; Pueblas, J.; Serrano, A.; Puente, R.; Corral, R. Adapting Artificial Neural Networks Training Algorithms to Adjoint-Based Aerodynamic Shape Optimization. In Proceedings of the Volume 13D: Turbomachinery—Multidisciplinary Design Approaches, Optimization, and Uncertainty Quantification, Boston, MA, USA, 26–30 June 2023; Radial Turbomachinery Aerodynamics. Unsteady Flows in Turbomachinery: Boston, MA, USA, 2023; p. V13DT34A021. [Google Scholar] [CrossRef]
- Xu, X.; Huang, X.; Bi, D.; Zhou, M. A Combined Artificial-Intelligence Aerodynamic Design Method for a Transonic Compressor Rotor Based on Reinforcement Learning and Genetic Algorithm. Appl. Sci. 2023, 13, 1026. [Google Scholar] [CrossRef]
- Ghosh, S.; Anantha Padmanabha, G.; Peng, C.; Andreoli, V.; Atkinson, S.; Pandita, P.; Vandeputte, T.; Zabaras, N.; Wang, L. Inverse Aerodynamic Design of Gas Turbine Blades Using Probabilistic Machine Learning. J. Mech. Des. 2021, 144, 021706. [Google Scholar] [CrossRef]
- Zhou, H.; Yu, K.; Luo, Q.; Luo, L.; Du, W.; Wang, S. Design Methods and Strategies for Forward and Inverse Problems of Turbine Blades Based on Machine Learning. J. Therm. Sci. 2022, 31, 82–95. [Google Scholar] [CrossRef]
- Ghosh, S.; Andreoli, V.; Padmanabha, G.A.; Peng, C.; Atkinson, S.; Pandita, P.; Vandeputte, T.; Zabaras, N.; Wang, L. A Probabilistic Machine Learning Framework for Explicit Inverse Design of Industrial Gas Turbine Blades. In Proceedings of the Volume 9B: Structures and Dynamics—Fatigue, Fracture, and Life Prediction; Probabilistic Methods; Rotordynamics; Structural Mechanics and Vibration, Virtual, Online, 7–11 June 2021; p. V09BT27A003. [Google Scholar] [CrossRef]
- Zeinalzadeh, A.; Pakatchian, M. Evaluation of Novel-Objective Functions in the Design Optimization of a Transonic Rotor by Using Deep Learning. Eng. Appl. Comput. Fluid Mech. 2021, 15, 561–583. [Google Scholar] [CrossRef]
- Bohara, R.P.; Linforth, S.; Thai, H.T.; Nguyen, T.; Ghazlan, A.; Ngo, T. Multi-Objective Bulk Scale Optimisation of an Auxetic Structure to Enhance Protection Performance. Eng. Struct. 2023, 280, 115729. [Google Scholar] [CrossRef]
- Nasrullah, A.I.H.; Puji Santosa, S.; Widagdo, D.; Arifurrahman, F. Structural Lattice Topology and Material Optimization for Battery Protection in Electric Vehicles Subjected to Ground Impact Using Artificial Neural Networks and Genetic Algorithms. Materials 2021, 14, 7618. [Google Scholar] [CrossRef]
- Liu, K.; Xu, Z.; Detwiler, D.; Tovar, A. Optimal Design of Cellular Material Systems for Crashworthiness. In Proceedings of the SAE 2016 World Congress and Exhibition, Detroit, MI, USA, 12–14 April 2016. SAE Technical Paper 2016-01-1396. [Google Scholar] [CrossRef]
- Liu, K.; Tovar, A.; Nutwell, E.; Detwiler, D. Thin-Walled Compliant Mechanism Component Design Assisted by Machine Learning and Multiple Surrogates. In Proceedings of the SAE 2015 World Congress & Exhibition, Detroit, MI, USA, 21–23 April 2015. SAE Technical Paper 2015-01-1369. [Google Scholar] [CrossRef]
- Choi, J.; Kim, H.; Noh, T.; Kang, Y.J.; Noh, Y. Digital Twin Data-Driven Multi-Disciplinary and Multi-Objective Optimization Framework for Automatic Design of Negative Stiffness Honeycomb. Int. J. Precis. Eng. Manuf. 2023, 24, 1453–1472. [Google Scholar] [CrossRef]
- Bucher, M.J.J.; Kraus, M.A.; Rust, R.; Tang, S. Performance-Based Generative Design for Parametric Modeling of Engineering Structures Using Deep Conditional Generative Models. Autom. Constr. 2023, 156, 105128. [Google Scholar] [CrossRef]
- Venugopal, V.; Anand, S. Structural and Thermal Generative Design Using Reinforcement Learning-Based Search Strategy for Additive Manufacturing. Manuf. Lett. 2023, 35, 564–575. [Google Scholar] [CrossRef]
- Almasri, W.; Danglade, F.; Bettebghor, D.; Adjed, F.; Ababsa, F. A Data-Driven Topology Optimization Approach to Handle Geometrical Manufacturing Constraints in the Earlier Steps of the Design Phase. Procedia CIRP 2023, 119, 377–383. [Google Scholar] [CrossRef]
- Kazemi, H.; Seepersad, C.C.; Alicia Kim, H. Multiphysics Design Optimization via Generative Adversarial Networks. J. Mech. Des. 2022, 144, 121702. [Google Scholar] [CrossRef]
- Fei, Y.; Liao, W.; Huang, Y.; Lu, X. Knowledge-Enhanced Generative Adversarial Networks for Schematic Design of Framed Tube Structures. Autom. Constr. 2022, 144, 104619. [Google Scholar] [CrossRef]
- Kallioras, N.A.; Lagaros, N.D. MLGen: Generative Design Framework Based on Machine Learning and Topology Optimization. Appl. Sci. 2021, 11, 12044. [Google Scholar] [CrossRef]
- Hertlein, N.; Buskohl, P.R.; Gillman, A.; Vemaganti, K.; Anand, S. Generative Adversarial Network for Early-Stage Design Flexibility in Topology Optimization for Additive Manufacturing. J. Manuf. Syst. 2021, 59, 675–685. [Google Scholar] [CrossRef]
- Almasri, W.; Bettebghor, D.; Ababsa, F.; Danglade, F.; Adjed, F. Deep Learning Architecture for Topological Optimized Mechanical Design Generation with Complex Shape Criterion. In Advances and Trends in Artificial Intelligence. Artificial Intelligence Practices; Fujita, H., Selamat, A., Lin, J.C.W., Ali, M., Eds.; Springer International Publishing: Cham, Switzerland, 2021; Volume 12798, pp. 222–234. [Google Scholar] [CrossRef]
- Oh, S.; Jung, Y.; Kim, S.; Lee, I.; Kang, N. Deep Generative Design: Integration of Topology Optimization and Generative Models. J. Mech. Des. 2019, 141, 111405. [Google Scholar] [CrossRef]
- Oh, S.; Jung, Y.; Lee, I.; Kang, N. Design Automation by Integrating Generative Adversarial Networks and Topology Optimization. In Proceedings of the Volume 2A: 44th Design Automation Conference, Quebec City, QC, Canada, 24–28 June 2018; p. V02AT03A008. [Google Scholar] [CrossRef]
- Antunes, A.P.; Azevedo, J.L.F. An Aerodynamic Optimization Computational Framework Using Genetic Algorithms. J. Braz. Soc. Mech. Sci. Eng. 2016, 38, 1037–1058. [Google Scholar] [CrossRef]
- Sun, G.; Sun, Y.; Wang, S. Artificial Neural Network Based Inverse Design: Airfoils and Wings. Aerosp. Sci. Technol. 2015, 42, 415–428. [Google Scholar] [CrossRef]
- Pehlivanoglu, Y.V. Direct and Indirect Design Prediction in Genetic Algorithm for Inverse Design Problems. Appl. Soft Comput. 2014, 24, 781–793. [Google Scholar] [CrossRef]
- Thinakaran, K.; Rajasekar, R. Design of Airfoil Using Backpropagation Training with Combined Approach. In Innovative Design, Analysis and Development Practices in Aerospace and Automotive Engineering; Bajpai, R.P., Chandrasekhar, U., Arankalle, A.R., Eds.; Springer: New Delhi, India, 2014; pp. 295–306. [Google Scholar] [CrossRef]
- Lee, S.; Lee, D.H.; Kim, K.H.; Park, T.C.; Lim, B.J.; Kang, Y.S. Multi-Disciplinary Design Optimization and Performance Evaluation of a Single Stage Transonic Axial Compressor. J. Mech. Sci. Technol. 2013, 27, 3309–3318. [Google Scholar] [CrossRef]
- Ulaganathan, S.; Asproulis, N. Surrogate Models for Aerodynamic Shape Optimisation. In Surrogate-Based Modeling and Optimization; Koziel, S., Leifsson, L., Eds.; Springer: New York, NY, USA, 2013; pp. 285–312. [Google Scholar] [CrossRef]
- Astrua, P.; Piola, S.; Silingardi, A.; Bonzani, F. Multi-Objective Constrained Aero-Mechanical Optimization of an Axial Compressor Transonic Blade. In Proceedings of the Volume 8: Turbomachinery, Parts A, B, and C, Copenhagen, Denmark, 11–15 June 2012; pp. 241–252. [Google Scholar] [CrossRef]
- Wang, X.; Hirsch, C.; Kang, S.; Lacor, C. Multi-Objective Optimization of Turbomachinery Using Improved NSGA-II and Approximation Model. Comput. Methods Appl. Mech. Eng. 2011, 200, 883–895. [Google Scholar] [CrossRef]
- Mengistu, T.; Ghaly, W. Aerodynamic Optimization of Turbomachinery Blades Using Evolutionary Methods and ANN-based Surrogate Models. Optim. Eng. 2008, 9, 239–255. [Google Scholar] [CrossRef]
- Zhou, C.; Zhang, X.; Yu, C. How Does AI Promote Design Iteration? The Optimal Time to Integrate AI into the Design Process. J. Eng. Des. 2023, 1–28. [Google Scholar] [CrossRef]
- Agius, S.; Farrugia, P.; Francalanza, E. Designing for Human Factors: Development and Evaluation of a Harmonistic Knowledge-Based Design Decision Support Tool. Proc. Des. Soc. 2023, 3, 1367–1376. [Google Scholar] [CrossRef]
- Aalaei, M.; Saadi, M.; Rahbar, M.; Ekhlassi, A. Architectural layout generation using a graph-constrained conditional Generative Adversarial Network (GAN). Autom. Constr. 2023, 155, 105053. [Google Scholar] [CrossRef]
- Li, K.; Ye, W. A Generative Adversarial Active Learning Method for Mechanical Layout Generation. Neural Comput. Appl. 2023, 35, 19315–19335. [Google Scholar] [CrossRef]
- Chen, R.; Zhao, J.; Yao, X.; Jiang, S.; He, Y.; Bao, B.; Luo, X.; Xu, S.; Wang, C. Generative Design of Outdoor Green Spaces Based on Generative Adversarial Networks. Buildings 2023, 13, 1083. [Google Scholar] [CrossRef]
- Zhang, C.; Tao, M.X.; Wang, C.; Fan, J.S. End-to-end Generation of Structural Topology for Complex Architectural Layouts with Graph Neural Networks. Comput.-Aided Civ. Infrastruct. Eng. 2024, 39, 756–775. [Google Scholar] [CrossRef]
- Ghasemi, P.; Yuan, C.; Marion, T.; Moghaddam, M. Are Generative Adversarial Networks Capable of Generating Novel and Diverse Design Concepts? An Experimental Analysis of Performance. Proc. Des. Soc. 2023, 3, 633–644. [Google Scholar] [CrossRef]
- Li, R.; Chi, H.L.; Peng, Z.; Chen, J. Tower Crane Layout Planning Through Generative Adversarial Network. In ECPPM 2022—eWork and eBusiness in Architecture, Engineering and Construction 2022, 1st ed.; CRC Press: London, UK, 2023; pp. 382–388. [Google Scholar] [CrossRef]
- Heyrani Nobari, A.; Rashad, M.F.; Ahmed, F. CreativeGAN: Editing Generative Adversarial Networks for Creative Design Synthesis. In Proceedings of the Volume 3A: 47th Design Automation Conference (DAC), Virtual, Online, 5–9 December 2021; p. V03AT03A002. [Google Scholar] [CrossRef]
- Kallioras, N.A.; Lagaros, N.D. DzAI: Deep Learning Based Generative Design. Procedia Manuf. 2020, 44, 591–598. [Google Scholar] [CrossRef]
- Brünnhäußer, J.; Lünnemann, P.; Bisang, U.; Novikov, R.; Flachmeier, F.; Wolff, M. Geometric Design Process Automation with Artificial Intelligence. In Advances in Production Management Systems. Smart Manufacturing and Logistics Systems: Turning Ideas into Action; Kim, D.Y., Von Cieminski, G., Romero, D., Eds.; Springer Nature: Cham, Switzerland, 2022; Volume 663, pp. 35–42. [Google Scholar] [CrossRef]
- Bertagna, F.; D’Acunto, P.; Ohlbrock, P.O.; Moosavi, V. Holistic Design Explorations of Building Envelopes Supported by Machine Learning. J. Facade Des. Eng. 2021, 9, 31–46. [Google Scholar]
- Fisher-Gewirtzman, D.; Polak, N. A Learning Automated 3D Architecture Synthesis Model: Demonstrating a Computer Governed Design of Minimal Apartment Units Based on Human Perceptual and Physical Needs. Archit. Sci. Rev. 2019, 62, 301–312. [Google Scholar] [CrossRef]
- Sato, H.; Igarashi, H. Fast Topology Optimization for PM Motors Using Variational Autoencoder and Neural Networks With Dropout. IEEE Trans. Magn. 2023, 59, 8200904. [Google Scholar] [CrossRef]
- Xu, Y.; Wang, B.; Sakamoto, Y.; Yamamoto, T.; Nishimura, Y.; Koike-Akino, T.; Wang, Y. Electric Machine Inverse Design with Variational Auto-Encoder (VAE). In Proceedings of the 2023 IEEE Energy Conversion Congress and Exposition (ECCE), Nashville, TN, USA, 29 October–2 November 2023; pp. 6338–6344. [Google Scholar] [CrossRef]
- Sato, H.; Igarashi, H. Visual Interpretation of Topology Optimization Results Based on Deep Learning. IEEE Trans. Magn. 2024, 60, 9200104. [Google Scholar] [CrossRef]
- Shigematsu, H.; Wakao, S.; Murata, N.; Makino, H.; Takeuchi, K.; Matsushita, M. Multi-Objective Topology Optimization of Synchronous Reluctance Motor Using Response Surface Approximation Derived by Deep Learning. IEEJ Trans. Electr. Electron. Eng. 2023, 18, 120–128. [Google Scholar] [CrossRef]
- Sasaki, H.; Hidaka, Y.; Igarashi, H. Prediction of IPM Machine Torque Characteristics Using Deep Learning Based on Magnetic Field Distribution. IEEE Access 2022, 10, 60814–60822. [Google Scholar] [CrossRef]
- Sasaki, H.; Hidaka, Y.; Igarashi, H. Explainable Deep Neural Network for Design of Electric Motors. IEEE Trans. Magn. 2021, 57, 8203504. [Google Scholar] [CrossRef]
- Doi, S.; Sasaki, H.; Igarashi, H. Multi-Objective Topology Optimization of Rotating Machines Using Deep Learning. IEEE Trans. Magn. 2019, 55, 7202605. [Google Scholar] [CrossRef]
- Sasaki, H.; Igarashi, H. Topology Optimization Accelerated by Deep Learning. IEEE Trans. Magn. 2019, 55, 7401305. [Google Scholar] [CrossRef]
- Sasaki, H.; Igarashi, H. Topology Optimization of IPM Motor with Aid of Deep Learning. Int. J. Appl. Electromagn. Mech. 2019, 59, 87–96. [Google Scholar] [CrossRef]
- Zhao, J.; Pei, J.; Yuan, J.; Wang, W. Structural Optimization of Multistage Centrifugal Pump via Computational Fluid Dynamics and Machine Learning Method. J. Comput. Des. Eng. 2023, 10, 1204–1218. [Google Scholar] [CrossRef]
- Agnolucci, A.; Marconcini, M.; Biliotti, D.; Toni, L.; Baroni, A.; Romani, L.; Arnone, A. Robust Design of a Vaned Diffuser in a Centrifugal Compressor Stage With a High Load Impeller. In Proceedings of the Volume 13D: Turbomachinery—Multidisciplinary Design Approaches, Optimization, and Uncertainty Quantification; Radial Turbomachinery Aerodynamics; Unsteady Flows in Turbomachinery, Boston, MA, USA, 26–30 June 2023; p. V13DT35A022. [Google Scholar] [CrossRef]
- Xie, X.; Li, Z.; Zhu, B.; Wang, H.; Zhang, W. Multi-Objective Optimization Design of a Centrifugal Impeller by Positioning Splitters Using GMDH, NSGA-III and Entropy Weight-TOPSIS. J. Mech. Sci. Technol. 2021, 35, 2021–2034. [Google Scholar] [CrossRef]
- Han, X.; Kang, Y.; Sheng, J.; Hu, Y.; Zhao, W. Centrifugal Pump Impeller and Volute Shape Optimization via Combined NUMECA, Genetic Algorithm, and Back Propagation Neural Network. Struct. Multidiscip. Optim. 2020, 61, 381–409. [Google Scholar] [CrossRef]
- Pei, J.; Gan, X.; Wang, W.; Yuan, S.; Tang, Y. Multi-Objective Shape Optimization on the Inlet Pipe of a Vertical Inline Pump. J. Fluids Eng. 2019, 141, 061108. [Google Scholar] [CrossRef]
- An, Z.; Zhounian, L.; Peng, W.; Linlin, C.; Dazhuan, W. Multi-Objective Optimization of a Low Specific Speed Centrifugal Pump Using an Evolutionary Algorithm. Eng. Optim. 2016, 48, 1251–1274. [Google Scholar] [CrossRef]
- Wang, W.; Pei, J.; Yuan, S.; Zhang, J.; Yuan, J.; Xu, C. Application of Different Surrogate Models on the Optimization of Centrifugal Pump. J. Mech. Sci. Technol. 2016, 30, 567–574. [Google Scholar] [CrossRef]
- Zhang, J.; Zhu, H.; Yang, C.; Li, Y.; Wei, H. Multi-Objective Shape Optimization of Helico-Axial Multiphase Pump Impeller Based on NSGA-II and ANN. Energy Convers. Manag. 2011, 52, 538–546. [Google Scholar] [CrossRef]
- Jeong, H.; Batuwatta-Gamage, C.; Bai, J.; Xie, Y.M.; Rathnayaka, C.; Zhou, Y.; Gu, Y. A Complete Physics-Informed Neural Network-based Framework for Structural Topology Optimization. Comput. Methods Appl. Mech. Eng. 2023, 417, 116401. [Google Scholar] [CrossRef]
- Yan, J.; Geng, D.; Xu, Q.; Li, H. Real-Time Topology Optimization Based on Convolutional Neural Network by Using Retrain Skill. Eng. Comput. 2023, 39, 4045–4059. [Google Scholar] [CrossRef]
- Ma, Q.; DeMeter, E.C.; Basu, S. Learning Topology Optimization Process via Convolutional Long-short-term Memory Autoencoder-decoder. Int. J. Numer. Methods Eng. 2023, 124, 2571–2588. [Google Scholar] [CrossRef]
- Seo, J.; Kapania, R.K. Development of Deep Convolutional Neural Network for Structural Topology Optimization. In Proceedings of the AIAA SCITECH 2022 Forum, San Diego, CA, USA, & Virtual, 3–7 January 2022. [Google Scholar] [CrossRef]
- Jeong, H.; Bai, J.; Batuwatta-Gamage, C.; Rathnayaka, C.; Zhou, Y.; Gu, Y. A Physics-Informed Neural Network-based Topology Optimization (PINNTO) Framework for Structural Optimization. Eng. Struct. 2023, 278, 115484. [Google Scholar] [CrossRef]
- Kushwaha, O.H.N. Topology Optimization Using Nonlinear Finite Element Analysis. Commun. Appl. Nonlinear Anal. 2023, 30, 17–28. [Google Scholar] [CrossRef]
- Tan, R.K.; Qian, C.; Li, K.; Xu, D.; Ye, W. An Adaptive and Scalable Artificial Neural Network-Based Model-Order-Reduction Method for Large-Scale Topology Optimization Designs. Struct. Multidiscip. Optim. 2022, 65, 348. [Google Scholar] [CrossRef]
- Hoang, V.N.; Nguyen, N.L.; Tran, D.Q.; Vu, Q.V.; Nguyen-Xuan, H. Data-Driven Geometry-Based Topology Optimization. Struct. Multidiscip. Optim. 2022, 65, 69. [Google Scholar] [CrossRef]
- Wang, D.; Xiang, C.; Pan, Y.; Chen, A.; Zhou, X.; Zhang, Y. A Deep Convolutional Neural Network for Topology Optimization with Perceptible Generalization Ability. Eng. Optim. 2022, 54, 973–988. [Google Scholar] [CrossRef]
- Esfahanian, V.; Izadi, M.J.; Bashi, H.; Ansari, M.; Tavakoli, A.; Kordi, M. Aerodynamic Shape Optimization of Gas Turbines: A Deep Learning Surrogate Model Approach. Struct. Multidiscip. Optim. 2024, 67, 2. [Google Scholar] [CrossRef]
- Wang, Y.; Shimada, K.; Barati Farimani, A. Airfoil GAN: Encoding and Synthesizing Airfoils for Aerodynamic Shape Optimization. J. Comput. Des. Eng. 2023, 10, 1350–1362. [Google Scholar] [CrossRef]
- Yi, J.; Deng, F. Cooperation of Thin-Airfoil Theory and Deep Learning for a Compact Airfoil Shape Parameterization. Aerospace 2023, 10, 650. [Google Scholar] [CrossRef]
- Sun, Y.; Sengupta, U.; Juniper, M. Physics-Informed Deep Learning for Simultaneous Surrogate Modeling and PDE-constrained Optimization of an Airfoil Geometry. Comput. Methods Appl. Mech. Eng. 2023, 411, 116042. [Google Scholar] [CrossRef]
- Shirvani, A.; Nili-Ahmadabadi, M.; Ha, M.Y. Machine Learning-Accelerated Aerodynamic Inverse Design. Eng. Appl. Comput. Fluid Mech. 2023, 17, 2237611. [Google Scholar] [CrossRef]
- Kim, M.I.; Yoon, H.S. Geometric Modification for the Enhancement of an Airfoil Performance Using Deep CNN. Ocean Eng. 2022, 266, 113000. [Google Scholar] [CrossRef]
- Zan, B.W.; Han, Z.H.; Xu, C.Z.; Liu, M.Q.; Wang, W.Z. High-Dimensional Aerodynamic Data Modeling Using a Machine Learning Method Based on a Convolutional Neural Network. Adv. Aerodyn. 2022, 4, 39. [Google Scholar] [CrossRef]
- Balla, K.; Sevilla, R.; Hassan, O.; Morgan, K. Inverse Aerodynamic Design Using Neural Networks. In Advances in Computational Methods and Technologies in Aeronautics and Industry; Knoerzer, D., Periaux, J., Tuovinen, T., Eds.; Springer International Publishing: Cham, Switzerland, 2022; Volume 57, pp. 131–143. [Google Scholar] [CrossRef]
- Lei, R.; Bai, J.; Wang, H.; Zhou, B.; Zhang, M. Deep Learning Based Multistage Method for Inverse Design of Supercritical Airfoil. Aerosp. Sci. Technol. 2021, 119, 107101. [Google Scholar] [CrossRef]
- Du, X.; He, P.; Martins, J.R. Rapid Airfoil Design Optimization via Neural Networks-Based Parameterization and Surrogate Modeling. Aerosp. Sci. Technol. 2021, 113, 106701. [Google Scholar] [CrossRef]
- Dai, S.; Li, Y.; Grace, K.; Globa, A. Towards Human-AI Collaborative Architectural Concept Design via Semantic AI. In Computer-Aided Architectural Design. INTERCONNECTIONS: Co-Computing Beyond Boundaries; Turrin, M., Andriotis, C., Rafiee, A., Eds.; Springer Nature: Cham, Switzerland, 2023; Volume 1819, pp. 68–82. [Google Scholar] [CrossRef]
- Wang, D.; Han, J. Exploring the Impact Of Generative Stimuli on the Creativity of Designers in Combinational Design. Proc. Des. Soc. 2023, 3, 1805–1814. [Google Scholar] [CrossRef]
- Yang, M.; Zang, T.; Jiang, P. Social Computational Design Method for Generating Product Shapes with GAN and Transformer Models. IFAC-PapersOnLine 2022, 55, 281–286. [Google Scholar] [CrossRef]
- Baum, E.B. On the Capabilities of Multilayer Perceptrons. J. Complex. 1988, 4, 193–215. [Google Scholar] [CrossRef]
- Goodfellow, I.; Pouget-Abadie, J.; Mirza, M.; Xu, B.; Warde-Farley, D.; Ozair, S.; Courville, A.; Bengio, Y. Generative Adversarial Nets. arXiv 2014, arXiv:1406.2661v1. [Google Scholar]
- Kingma, D.P.; Welling, M. Auto-Encoding Variational Bayes. arXiv 2022, arXiv:1312.6114. [Google Scholar] [CrossRef]
- Hochreiter, S.; Schmidhuber, J. Long Short-Term Memory. Neural Comput. 1997, 9, 1735–1780. [Google Scholar] [CrossRef]
- Raissi, M.; Perdikaris, P.; Karniadakis, G. Physics-Informed Neural Networks: A Deep Learning Framework for Solving Forward and Inverse Problems Involving Nonlinear Partial Differential Equations. J. Comput. Phys. 2019, 378, 686–707. [Google Scholar] [CrossRef]
- Bengio, Y.; Ducharme, R.; Vincent, P. A Neural Probabilistic Language Model. J. Mach. Learn. Res. 2003, 3, 1137–1155. [Google Scholar]
- Szepesvári, C. Algorithms for Reinforcement Learning; Morgan and Claypool: San Rafael, CA, USA, 2010; Volume 4. [Google Scholar] [CrossRef]
- Wang, M.Y.; Wang, X.; Guo, D. A Level Set Method for Structural Topology Optimization. Comput. Methods Appl. Mech. Eng. 2003, 192, 227–246. [Google Scholar] [CrossRef]
- Man, K.F.; Tang, K.S.; Kwong, S. Genetic Algorithms: Concepts and Applications [in Engineering Design]. IEEE Trans. Ind. Electron. 1996, 43, 519–534. [Google Scholar] [CrossRef]
- Chen, J.; Shapiro, V.; Suresh, K.; Tsukanov, I. Shape Optimization with Topological Changes and Parametric Control. Int. J. Numer. Methods Eng. 2007, 71, 313–346. [Google Scholar] [CrossRef]
- Arjovsky, M.; Chintala, S.; Bottou, L. Wasserstein GAN. arXiv 2017, arXiv:1701.07875. [Google Scholar] [CrossRef]
- Rade, J.; Balu, A.; Herron, E.; Pathak, J.; Ranade, R.; Sarkar, S.; Krishnamurthy, A. Algorithmically-Consistent Deep Learning Frameworks for Structural Topology Optimization. Eng. Appl. Artif. Intell. 2021, 106, 104483. [Google Scholar] [CrossRef]
- Bielecki, D.; Patel, D.; Rai, R.; Dargush, G.F. Multi-Stage Deep Neural Network Accelerated Topology Optimization. Struct. Multidiscip. Optim. 2021, 64, 3473–3487. [Google Scholar] [CrossRef]
- Veličković, P.; Vitvitskyi, A.; Markeeva, L.; Ibarz, B.; Buesing, L.; Balog, M.; Novikov, A. Amplifying Human Performance in Combinatorial Competitive Programming. arXiv 2024, arXiv:2411.19744. [Google Scholar] [CrossRef]
- Chang, C.C.; Shen, Y.; Fan, S.; Li, J.; Zhang, S.; Cao, N.; Chen, Y.; Zhang, X. LaMAGIC: Language-Model-Based Topology Generation for Analog Integrated Circuits. In Proceedings of the 41st International Conference on Machine Learning, ICML’24, Vienna, Austria, 21–27 July 2025; Volume 235, pp. 6253–6262. [Google Scholar]
- Novikov, A.; Vu, N.; Eisenberger, M.; Dupont, E.; Huang, P.S.; Wagner, A.Z.; Shirobokov, S.; Kozlovskii, B.; Ruiz, F.J.R.; Mehrabian, A.; et al. AlphaEvolve: A Coding Agent for Scientific and Algorithmic Discovery. arXiv 2025, arXiv:2506.13131. [Google Scholar] [CrossRef]
- Wu, S.; Khasahmadi, A.; Katz, M.; Jayaraman, P.K.; Pu, Y.; Willis, K.; Liu, B. CAD-LLM: Large Language Model for CAD Generation. In Proceedings of the 37th Conference on Neural Information Processing Systems, New Orleans, LA, USA, 10–16 December 2023. [Google Scholar]
- Wu, S.; Khasahmadi, A.; Katz, M.; Jayaraman, P.K.; Pu, Y.; Willis, K.; Liu, B. CadVLM: Bridging Language and Vision in the Generation of Parametric CAD Sketches. arXiv 2024, arXiv:2409.17457. [Google Scholar] [CrossRef]
- Herron, E.; Rade, J.; Jignasu, A.; Ganapathysubramanian, B.; Balu, A.; Sarkar, S.; Krishnamurthy, A. Latent Diffusion Models for Structural Component Design. Comput.-Aided Des. 2024, 171, 103707. [Google Scholar] [CrossRef]
- Xu, X.; Lambourne, J.G.; Jayaraman, P.K.; Wang, Z.; Willis, K.D.D.; Furukawa, Y. BrepGen: A B-rep Generative Diffusion Model with Structured Latent Geometry. arXiv 2024, arXiv:2401.15563. [Google Scholar] [CrossRef]
- Sato, H.; Igarashi, H. Deep Learning-Based Surrogate Model for Fast Multi-Material Topology Optimization of IPM Motor. COMPEL— Int. J. Comput. Math. Electr. Electron. Eng. 2022, 41, 900–914. [Google Scholar] [CrossRef]
- Kumar, A.; Vadlamani, N.R. Inverse Design of Airfoils Using Convolutional Neural Network and Deep Neural Network. In Proceedings of the ASME 2021 Gas Turbine India Conference, Virtual, Online, 2–3 December 2021; p. V001T02A001. [Google Scholar] [CrossRef]
- Khan, S.; Goucher-Lambert, K.; Kostas, K.; Kaklis, P. ShipHullGAN: A Generic Parametric Modeller for Ship Hull Design Using Deep Convolutional Generative Model. Comput. Methods Appl. Mech. Eng. 2023, 411, 116051. [Google Scholar] [CrossRef]
- Tan, Z.; Li, R.; Zhang, Y. Flow Field Reconstruction of 2D Hypersonic Inlets Based on a Variational Autoencoder. Aerospace 2023, 10, 825. [Google Scholar] [CrossRef]
- Chen, L.; Yuan, Z.; Ji, H.; Sun, L.; Peter, C.; Wang, B. A Foundation Model Enhanced Approach for Generative Design in Combinational Creativity. J. Eng. Des. 2024, 35, 1394–1420. [Google Scholar] [CrossRef]
- Wu, Y.; Kouta, M.; Yun Suen, P. OwnDiffusion: A Design Pipeline Using Design Generative AI to Preserve Sense of Ownership. In Proceedings of the SIGGRAPH Asia 2023 Posters, Sydney, Australia, 12–15 December 2023; pp. 1–2. [Google Scholar] [CrossRef]
- Wang, Z.; Tan, Z.; Ma, Y. Combinediff: A GenAI Creative Support Tool for Image Combination Exploration. J. Eng. Des. 2025, 36, 731–757. [Google Scholar] [CrossRef]
- Chen, L.; Cai, Z.; Cheang, W.; Long, Q.; Sun, L.; Childs, P.; Zuo, H. AskNatureGPT: An LLM-Driven Concept Generation Method Based on Bio-Inspired Design Knowledge. J. Eng. Des. 2025, 1–35. [Google Scholar] [CrossRef]
- Hu, X.; Ang, L.; Dai, Y. Combining ChatGPT and Knowledge Graph for Explainable Machine Learning-Driven Design: A Case Study. J. Eng. Des. 2025, 1–23. [Google Scholar] [CrossRef]
- Foster, T.; Foerster, J. Learning to Reason at the Frontier of Learnability. arXiv 2025, arXiv:2502.12272. [Google Scholar]
- McInnes, L.; Healy, J.; Melville, J. UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction. arXiv 2020, arXiv:1802.03426. [Google Scholar] [CrossRef]
- Autodesk. Generative Design; Autodesk: San Francisco, CA, USA, 2024. [Google Scholar]
Review Paper | Year | Focus | Method |
---|---|---|---|
Regenwetter et al. [19] | 2022 | Deep generative models in engineering design | Initial keyword search for GAN, VAE, and RL at select venues yielded 41 papers, expanded to 63 through citation tracking. Analysis structured by algorithms, datasets, representations, and applications. |
Khanolkar et al. [20] | 2023 | AI integration in engineering design process | Keyword search. Examined 108 peer-reviewed studies. Structured by engineering design stages. |
Kretzschmar et al. [21] | 2024 | Generative AI in engineering and product development | Keyword search. Title and abstract filtering. Examined 28 studies. |
This review | 2025 | Intersection of AI and GD | Keyword search. NLP based topic filtering. Examined 88 studies. Structured by technique, function, application, benefits, limitations, and future work. |
Attribute | Definition | Example |
---|---|---|
AI Techniques | Specific machine learning and AI models used to process data and generate designs. | Multilayer Perceptron (MLP) |
GD Techniques | Approaches used to create, optimise, or explore design solutions computationally. | Density Methods |
AI Function | The role performed by a technique in the context of the study method. | Surrogate Model |
Application | The specific field or industry where AI and GD methods are implemented. | Aerodynamic Design |
Benefits | Positive outcomes achieved by using AI and GD, such as faster or more efficient designs. | Promotes Design Diversity |
Limitations | Drawbacks or challenges associated with using AI and GD methods in practice. | Data and Training Costs |
Future Work | Stated goals for future development of the methods in the papers. | Extend to 3D Environment |
Publication Source | # |
---|---|
Lecture Notes in Computer Science | 159 |
IEEE Access | 138 |
Journal of Computational Design and Engineering | 137 |
Scientific Reports | 118 |
Proceedings of the ASME Design Engineering Technical Conference | 115 |
Topic | References | |
---|---|---|
1 | ML and AI in Aerodynamic Design and Optimisation | [32,33,34,35,36,37,38,39] |
2 | AI-Assisted Multi-Objective Optimisation in Structural Engineering Design | [40,41,42,43,44] |
3 | GenAI and ML for Engineering Design and Materials Optimisation | [45,46,47,48,49,50,51,52,53,54] |
4 | ML and Surrogate Modelling Techniques in Aerodynamic Design and Optimisation | [55,56,57,58,59,60,61,62,63] |
5 | Human-AI Collaboration in Design and Creative Processes | [64,65] |
6 | GenAI and DL for Automated Architectural and Structural Design | [66,67,68,69,70,71,72,73] |
7 | AI-Driven GD and Automation in Architecture and Engineering | [20,74,75,76] |
8 | ML-Driven Optimisation and Design of Electric Machines and Electromagnetic Devices | [5,77,78,79,80,81,82,83,84,85] |
9 | AI-Driven Multi-Objective Optimisation in Fluid Machinery Design | [23,86,87,88,89,90,91,92,93] |
10 | DL-Driven Innovations in TO | [3,94,95,96,97,98,99,100,101,102] |
11 | DL and ML Techniques for Aerodynamic Shape Design, Inverse Modelling, and Optimization | [103,104,105,106,107,108,109,110,111,112] |
12 | GenAI and Human-AI Collaboration in Creative Design | [113,114,115] |
Publication Source | # |
---|---|
Proceedings of the ASME Turbo Expo | 5 |
Structural and Multidisciplinary Optimisation | 4 |
Proceedings of the Design Society | 4 |
Automation in Construction | 3 |
Computer Methods in Applied Mechanics and Engineering | 3 |
Code | Diagram | Description | # | References in Final Corpus |
---|---|---|---|---|
MLP [116] | Multilayer Perceptron: A feedforward artificial neural network characterised by an input layer, a number of hidden layers, and an output layer. Each neuron in one layer is fully connected to every neuron in the subsequent layer via weighted connections. Non-linear activation functions applied at each neuron enable the network to learn non-linear relationships between inputs and outputs, making it suitable for classification and regression tasks. Training involves adjusting weights, usually via backpropagation. | 32 | [3,33,36,38,39,40,41,55,56,57,58,59,61,62,63,64,74,75,77,86,87,88,89,90,91,92,93,99,101,110,112,113] | |
GAN [117] | Generative Adversarial Network: A framework consisting of two neural networks—a Generator and a Discriminator—that are trained through adversarial competition. The Generator creates data whilst the Discriminator distinguishes between synthetic and real inputs. As training progresses, the Generator becomes capable of producing increasingly realistic outputs. | 22 | [47,48,49,51,52,53,54,64,66,67,68,69,70,71,72,104,105,111,112,114,115] | |
VAE [118] | Variational Auto-Encoder: A generative neural network model that learns a latent representation of input data. It consists on an Encoder, which maps input data into the latent space, and a Decoder, which samples the latent space to reconstruct samples. VAEs are trained to minimise reconstruction error. This also allows them to generate new data by sampling the learned latent space. | 5 | [5,45,77,78,104] | |
CNN [13] | Convolutional Neural Network: A specialised type of neural network designed to process 2D images. CNNs use convolutional layers to apply filters across the input data to detect local patterns. The convolutional layers are followed by pooling layers, which reduce dimensionality and MLP-like fully connected layers for classification or regression. They are used to learn hierarchical representations of features. | 18 | [32,50,52,79,80,81,82,83,84,85,95,97,100,102,103,108,109,111] | |
LSTM [119] | Long Short-Term Memory: An advanced type of neural network designed to remember long-range dependencies in sequential data. They are effective for time-series-based tasks. | 4 | [50,96,107,112] | |
PINN [120] | Physics Informed Neural Network: A type of neural network that contains embedded knowledge of physical laws (typically PDEs). PINNs are trained to minimise a composite objective function by minimising both data error and the PDE residual. They are used to ensure that predictions made about systems are physically consistent. | 3 | [94,98,106] | |
PM [121] | Probabilistic Model: A category of models that uses probability distributions for modelling uncertainty and making inferences. | 3 | [35,60,115] | |
RL [122] | Reinforcement Learning: A machine learning approach characterised by training via interactions between the agent and an environment. The agent takes actions in the environment and is rewarded or punished accordingly. With the goal to maximise reward, the agent learns the actions required to achieve its goal. The technique is often used for its adaptability in dynamic environments. | 3 | [20,35,46] | |
Other | Niche approaches not in standard categories. | 16 | [23,34,36,37,42,43,44,57,65,69,72,73,75,76,101,113] |
Code | Description | # | References in Final Corpus |
---|---|---|---|
AI | Often referred to as ‘direct AI generation’ in the text of this publication, this code captures a GD approach where an AI model directly produces a design representation (e.g., an image, voxel grid, or sets of parameters) as its primary output. This output requires minimal or no subsequent filtering or post-processing before being considered viable. This approach automates the ideation or form-finding phase by learning from example data. | 35 | [36,37,38,45,47,48,49,52,56,58,61,64,66,67,68,69,70,71,72,74,78,94,96,97,98,102,104,105,106,108,111,112,113,114,115] |
Density Methods | A class of TO where the design domain is discretised into finite elements and the material density within each element is treated as a design variable. The algorithm iteratively adjusts the densities to minimise an objective function (e.g., minimise compliance) subject to constraints (e.g., volume fraction). SIMP is the most prominent example of a density method of topology optimisation. It was first described in [16]. | 12 | [3,42,46,50,51,53,54,73,95,98,99,100] |
Level Set | A TO technique where the boundary of the design is represented implicitly as the zero-level set of a higher-dimensional scalar field. The scalar field is evolved with time to minimise the objective function. The implicit representation allows for complex topological changes [123]. | 1 | [80] |
GA | A population-based optimisation algorithm that simulates natural selection to iteratively evolve design solutions. Population members mutate and cross over, and the better-performing designs are propagated to converge toward workable solutions. The process is described in [124]. | 33 | [5,23,32,33,35,40,41,43,44,55,57,59,60,62,63,65,76,77,81,82,83,84,85,86,88,89,90,91,92,93,103,104,109] |
Parametric Shape Optimisation | An optimisation technique that refines geometry by adjusting its design parameters/variables. An example of the technique is given in [125]. | 4 | [34,39,79,107] |
Other | Encompasses miscellaneous or hybrid optimization techniques not categorized above. | 8 | [5,20,39,75,87,88,101,110] |
Code | Description | # | References in Final Corpus |
---|---|---|---|
Design Generation | The AI directly synthesises complete design outcomes. | 14 | [45,54,64,67,68,72,77,78,80,104,111,113,114,115] |
Provides Design Variation | The AI is used specifically to provide a wide range of design outcomes, often with the intention of reaching novel solutions. | 5 | [53,54,72,80,104] |
Optimisation | Uses AI to replace some or all of the steps usually done by an optimiser. | 30 | [5,20,34,36,37,38,46,47,48,49,50,51,52,56,57,58,66,69,71,73,74,76,78,94,96,101,102,105,108,115] |
Surrogate model | Replaces numerical physics solvers with an AI approximation [120]. | 43 | [23,32,33,35,38,39,40,41,42,44,55,59,60,61,62,63,65,77,78,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,97,98,99,103,106,107,109,110,112] |
Super-resolution | Upsamples and enhances discretised domains. | 1 | [100] |
Clusters the Design Space | Reduces the dimensionality of the design space by grouping solutions. | 3 | [42,43,75] |
Other | Various niche uses for AI. | 6 | [3,65,70,76,79,82] |
Code | Description | # | References in Final Corpus |
---|---|---|---|
Aerodynamic Design | Design and optimisation of fluid-interacting surfaces such as aerofoils and turbine blades. The goals often involve maximising lift or minimising drag. The process is often applied in aircraft and automotive design. | 26 | [32,33,35,36,37,38,39,55,56,57,58,59,60,61,62,63,103,104,105,106,107,108,109,110,111,112] |
Architectural Design | Encompasses the design of buildings, focussing on their form, layout and functionality. Requires engineering consideration of space planning, structural integrity and facade design. | 5 | [66,69,75,76,113] |
Pump Design | Focusses on the design of mechanical devices used to move fluids. Objectives include attaining specific flow rates, pressure heads and efficiencies. | 9 | [23,86,87,88,89,90,91,92,93] |
Motor Design | Involves the design and optimisation of electromagnetic components within electrical machines. Often focusses upon maximising torque and efficiency. | 9 | [5,77,78,79,80,81,82,83,85] |
Structural Design | Focusses on the design of components to efficiently and safely support loads. Objectives include minimising material usage and specific resonant frequencies. | 27 | [3,40,41,42,43,44,45,46,47,49,50,51,52,53,54,73,74,84,94,95,96,97,98,99,100,101,102] |
Layout Design | Aims to determine the optimal organisation of components within an area. Applied in domains such as office and PCB design the field seeks to maximise space utilisation. | 3 | [67,68,71] |
Other | Niche and varied application domains. | 9 | [20,34,48,64,65,70,72,114,115] |
Code | Description | # | References in Final Corpus |
---|---|---|---|
Speed | The application of AI enabled shorter design generation times. | 67 | [20,23,32,33,34,35,36,37,38,39,40,41,43,44,45,47,49,50,51,52,55,56,57,58,59,60,61,62,63,64,66,67,68,69,71,73,74,75,77,80,81,83,84,85,86,87,89,90,91,92,93,95,96,97,98,99,100,101,102,103,104,106,107,109,110,112,113] |
Improved Performance | The generated outcomes performed better (achieved higher objective function values). | 6 | [35,42,65,72,76,108] |
Generalises | The study’s method is applicable across a large range of scenarios. | 1 | [102] |
Interpretability | The justification behind why a specific outcome was reached is clear. | 3 | [43,75,79] |
Promotes Design Diversity | Ensures large design space exploration. | 12 | [46,50,53,54,64,67,68,78,80,104,113,114] |
Code | Description | # | References in Final Corpus |
---|---|---|---|
Data and Training Costs | The time and monetary costs of the AI training process was high. | 50 | [5,32,33,35,36,37,38,39,40,41,44,45,47,50,52,53,54,55,57,59,60,62,64,65,67,68,73,74,75,77,84,85,86,87,90,91,92,93,97,99,102,104,106,107,108,109,110,111,112,113] |
Accuracy and Stability | Inability of the AI to provide consistently physically valid outcomes. | 28 | [20,34,37,39,41,43,44,49,52,53,55,62,63,71,74,80,83,84,85,91,96,98,100,101,102,104,110,111] |
Optimisation and Convergence | Lack of reliable optimisation convergence. | 6 | [50,54,56,76,86,111] |
Generalisation and Repeatability | The ability of the technique to reliably operate in scenarios that it was not trained on. | 13 | [23,36,38,41,43,45,51,58,59,60,61,95,107] |
Data Quality | Poor training data not representative of the desired training function. | 15 | [5,33,55,56,57,64,68,73,75,93,97,99,102,106,112] |
Parametrisation | The use of AI limited the design space due to parametrisation. | 3 | [48,69,78] |
Black box | Generated outcomes are unexplainable. | 4 | [20,51,95,113] |
Training and Implementation Difficulty | A high degree of expertise is required to apply the techniques. | 5 | [36,47,53,68,98] |
N/A | Studies which did not discuss the limitations of their chosen AI technique. | 20 | [3,20,42,46,49,66,70,72,75,76,79,81,82,88,89,94,103,105,114,115] |
Code | Description | # | References in Final Corpus |
---|---|---|---|
Algorithmic Enhancement | The study stated an aim to improve the efficacy or speed of their method. | 33 | [3,20,23,32,33,36,39,42,44,45,50,54,56,61,64,66,67,68,69,70,74,75,76,85,95,102,107,108,109,111,112,113,114] |
Application and Domain Expansion | The desire to evaluate the developed method in new domains. | 21 | [3,5,37,38,40,41,45,49,50,51,55,72,77,79,83,85,94,100,104,111,115] |
Extend to 3D Environment | The study aims to develop a method from 2D into 3D. | 12 | [37,38,47,48,53,66,73,96,97,98,100,101] |
New Constraints and Objectives | The goal of testing the method with altered optimisation parameters. | 22 | [32,36,43,46,51,52,62,66,67,71,72,73,74,76,80,81,84,96,97,106,111,113] |
Validation and Testing | The aim of further investigating the existing method using new case studies to test generalisability. | 7 | [36,40,44,47,53,64,92] |
Manufacturability | The study wishes to evaluate and improve the physical feasibility of outcome designs. | 3 | [42,52,65] |
N/A | The study stated no aims to conduct further work. | 20 | [34,35,57,58,59,60,63,78,82,86,87,88,89,90,91,93,99,103,105,110] |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Peckham, O.; Raines, J.; Bulsink, E.; Goudswaard, M.; Gopsill, J.; Barton, D.; Nassehi, A.; Hicks, B. Artificial Intelligence in Generative Design: A Structured Review of Trends and Opportunities in Techniques and Applications. Designs 2025, 9, 79. https://doi.org/10.3390/designs9040079
Peckham O, Raines J, Bulsink E, Goudswaard M, Gopsill J, Barton D, Nassehi A, Hicks B. Artificial Intelligence in Generative Design: A Structured Review of Trends and Opportunities in Techniques and Applications. Designs. 2025; 9(4):79. https://doi.org/10.3390/designs9040079
Chicago/Turabian StylePeckham, Owen, Jonathan Raines, Erik Bulsink, Mark Goudswaard, James Gopsill, David Barton, Aydin Nassehi, and Ben Hicks. 2025. "Artificial Intelligence in Generative Design: A Structured Review of Trends and Opportunities in Techniques and Applications" Designs 9, no. 4: 79. https://doi.org/10.3390/designs9040079
APA StylePeckham, O., Raines, J., Bulsink, E., Goudswaard, M., Gopsill, J., Barton, D., Nassehi, A., & Hicks, B. (2025). Artificial Intelligence in Generative Design: A Structured Review of Trends and Opportunities in Techniques and Applications. Designs, 9(4), 79. https://doi.org/10.3390/designs9040079