Search

In creativity theory, combining two unrelated concepts into a novel idea is a common means of enhancing creativity. Designers can integrate the Additive concept into the Base concept to inspire and facilitate creative tasks. However, conceiving high-quality combinational ideas poses a challenge that combinational creativity itself demands the consideration of conceptual reasoning and synthesis. We propose an AI foundation model enhanced approach for supporting combinational creativity. This approach derives combinational embodiments, and assists humans in verbalising and externalising combinational ideas. Our experimental study demonstrates that the generated combinational ideas by the approach obtained highest scores compared to those ideas generated without an AI foundation model or combinational strategy. We built a combinational creativity tool called CombinatorX based on this approach to generate ideas. In a study with the comparison of an existing combinational creativity tool and Internet search, we validated that our approach improves the effectiveness of combinational idea generation, enables a reduction in labour force, and facilitates the refinement of combinational ideation. [ABSTRACT FROM AUTHOR]

In recent years, the field of structural DNA nanotechnology has advanced rapidly due to transformative design tools. Although these tools have been revolutionary, they still bear one overall limitation of requiring users to fully conceptualize their designs before designing. Recently, a simple computational casting technique was developed using generative optimization strategies to automate the DNA origami nanostructure design. This approach employs a shape annealing algorithm, which creates a formal language of honeycomb nanostructures with shape grammars and drives designs from the language toward a desired configuration using simulated annealing. This initial demonstration of the approach can generate novel scaffold routing schemes for creating solid or hollow structures constrained by the boundaries of polyhedral meshes. The results from the initial approach, particularly from the hollow structures, reveal a challenging design space. This simple technique generates novel scaffold routing schemes that do not replicate the overall polyhedral mesh shape and are limited in their ability to control scaffold path exploration in the design space. This paper demonstrates an approach for achieving different levels of consistent effective wall thicknesses and improving the quality of mesh coverage for hollow structures that can be tuned and optimized by introducing a more refined computational casting technique. We achieve these improvements through changes in the simulated annealing algorithm by adding a Hustin move set algorithm that dynamically adjusts the performance of the overall design and redefining how these hollow designs are articulated. This work illustrates how the technique can navigate a challenging design space to generate effective hollow designs. [ABSTRACT FROM AUTHOR]

This study presents a novel design methodology to automate the development process of a regolith habitation shell for a crewed mission on Mars. The integration of generative design, simulation tools, and data science methodologies allows for a systematic and comprehensive iterative design process, which facilitates the development of dual-habitation shells that can resist harsh conditions on Mars. The study establishes a methodology consisting of four fundamental sections. Initially, the environmental conditions on the surface of Mars are assessed and quantified as structural loads, so that the generative design tools effectively transform these stresses into design parameters that might be used in finite element modeling (FEM). Next, this study employs human factors derived from existing literature as input for Social Network Analysis (SNA) and the Label Propagation Algorithm (LPA) to address important design inquiries concerning functional zoning, layout configuration, geometry generation, and habitability variables. These inquiries aim to inform the decision-making process regarding autonomous typology design. Furthermore, the use of space syntax, physics simulations, and Voronoi algorithms improves the accuracy of the habitation planning process by generating floor designs and habitable volumes that align with the distinctive requirements and preferences of individuals. Lastly, the regolith habitation shell is constructed using structural simulation techniques that incorporate Mars' environmental pressures as design considerations. By employing Karamba3D and numerical finite element modeling (FEM) techniques, the structural performance of the regolith habitation shell under the environmental conditions on Mars is simulated. Consequently, this study introduces three distinct habitation shell designs that were developed and exposed to structural simulations to evaluate the effects of environmental stresses on Mars. The study explores the conjunction between generative design, data science, and structural simulations by employing a systematic and comprehensive design process. Its objective is to provide insights into the impact of environmental pressures on habitation shells and improve our understanding of design outcomes for future exploration on Mars. [ABSTRACT FROM AUTHOR]

Bridging the gaps among various categories of stochastic microstructures remains a challenge in the design representation of microstructural materials. Each microstructure category requires certain unique mathematical and statistical methods to define the design space (design representation). The design representation methods are usually incompatible between two different categories of stochastic microstructures. The common practice of preselecting the microstructure category and the associated design representation method before conducting rigorous computational design restricts the design freedom and hinders the discovery of innovative microstructure designs. To overcome this issue, this article proposes and compares two novel methods, the deep generative modeling-based method, and the curvature functional-based method, to understand their pros and cons in designing mixed-category stochastic microstructures for desired properties. For the deep generative modeling-based method, the variational autoencoder is employed to generate an unstructured latent space as the design space. For the curvature functional-based method, the microstructure geometry is represented by curvature functionals, of which the functional parameters are employed as the microstructure design variables. Regressors of the microstructure design variables-property relationship are trained for microstructure design optimization. A comparative study is conducted to understand the relative merits of these two methods in terms of computational cost, continuous transition, design scalability, design diversity, dimensionality of the design space, interpretability of the statistical equivalency, and design performance. [ABSTRACT FROM AUTHOR]

Computer-aided design (CAD) is a standard design tool used in engineering practice and by students. CAD has become increasingly analytic and inventive in incorporating artificial intelligence (AI) approaches to design, e.g., generative design (GD), to help expand designers' divergent thinking. However, generative design technologies are relatively new, we know little about generative design thinking in students. This research aims to advance our understanding of the relationship between aspects of generative design thinking and traditional design thinking. This study was set in an introductory graphics and design course where student designers used Fusion 360 to optimize a bicycle wheel frame. We collected the following data from the sample: divergent and convergent psychological tests and an open-ended response to a generative design prompt (called the generative design reasoning elicitation problem). A Spearman's rank correlation showed no statistically significant relationship between generative design reasoning and divergent thinking. However, an analysis of variance found a significant difference in generative design reasoning and convergent thinking between groups with moderate GD reasoning and low GD reasoning. This study shows that new computational tools might present the same challenges to beginning designers as conventional tools. Instructors should be aware of informed design practices and encourage students to grow into informed designers by introducing them to new technology, such as generative design. [ABSTRACT FROM AUTHOR]

Parametric modeling and generative deep learning for bridge design Given the significant environmental impact of the construction industry, the analysis and, above all, optimization of the sustainability of structures while maintaining established levels of reliability are becoming increasingly important. While there are initial tools for life cycle analysis in the building sector, these are not directly transferable to bridge structures. This paper introduces a deep learning‐based approach to explore and model the design space of parametric bridge models and their performance evaluations, illustrating its application for multi‐objective optimization of reinforced concrete frame bridges. Initially, data is synthetically generated using a parametric bridge model and coupling analysis software, followed by training a conditional variational autoencoder (CVAE) as a metamodel. In the context of conceptual bridge design, the CVAE serves as an efficient co‐pilot for both forward and inverse analysis. The sensitivity analysis performed with the CVAE reveals relationships between design parameters and/or performance metrics, highlighting optimization potentials. The integrated framework presented here has the potential to realize efficient bridge design, particularly focusing on sustainability and structural safety criteria, and can be easily extended to other parametric inquiries. [ABSTRACT FROM AUTHOR]

The use of machine learning in aerodynamic shape optimization problems has significantly increased in recent years. While existing deep learning techniques enable efficient design space exploration on data with an underlying Euclidean or grid-like structure, the direct optimization of non-parametric 3D geometries is still limited. In this article, we propose a geometric deep learning model that generates triangled-based meshed surfaces through the use of a graph variational autoencoder that learns the latent representations of a non-parametric 3D dataset. Once this framework is trained to embed all the input meshes in a properly distributed latent space, its exploration is managed by a genetic algorithm. In this regard, the NSGA-II is the agent in charge of sampling geometries that combine topology and aerodynamic features of the initial ones. Furthermore, in each iteration, it evaluates their aerodynamic performance with CFD in order to guide the optimization process and find the most effective region of the latent space. As a result, those solutions that maximize aerodynamic performance are provided through a Pareto front. The application to a case study and a real-world application is introduced aiming to validate the proposed approach. [ABSTRACT FROM AUTHOR]

As virtual worlds continue to rise in popularity, so do the expectations of users for the content of virtual scenes. Virtual worlds must be large in scope and offer enough freedom of movement to keep the audience occupied at all times. For content creators, it is difficult to keep up by manually producing the surrounding content. Therefore, the application of procedural modelling techniques is required. Virtual worlds often mimic the real world, which is composed of organized and connected outdoor and indoor layouts. It is expected that all content is present on the virtual scene and that a user can navigate streets, enter buildings, and interact with furniture within a single virtual world. While there are many procedural methods for generating different layout types, they mostly focus only on one layout type, whereas complete scene generation is greatly underrepresented. This paper aims to identify the coverage of layout types by different methods because similar issues exist for the generation of content of different layout types. When creating a new method for layout generation, it is important to know if the results of existing methods can be appended to other methods. This paper presents a survey of existing procedural modelling methods, which were organized into five categories based on the core approach: pure subdivision, grammar‐based, data‐driven, optimization, and simulation. Information about the covered layout types, the possibility of user interaction during the generation process, and the input and output shape types of the generated content is provided for each surveyed method. The input and output shape types of the generated content can be useful to identify which methods can continue the generation by using the output of other methods as their input. It was concluded that all surveyed methods work for only a few different layout types simultaneously. Moreover, only 35% of the surveyed methods offer interaction with the user after completing the initial process of space generation. Most existing approaches do not perform transformations of shape types. A significant number of methods use the irregular shape type as input and generate the same shape type as the output, which is sufficient for coverage of all layout types when generating a complete virtual world. [ABSTRACT FROM AUTHOR]

Inadequate highway design is a leading cause of traffic accidents, underscoring the importance of adhering to guidelines and regulations for highway design. These standards exist to safeguard road users by addressing crucial factors, like road geometry, signage, and lane markings. Thus, emphasis is placed on computational methods that can optimize towards higher levels of safety, capacity, efficiency, and sustainability in highway designs. Building Information Modeling (BIM) enhances this process by creating a digital model with physical and operational attributes. In this study, a user-friendly, logic-based language is utilized to encode rules for designing highway passing lanes by which designs are automatically evaluated and generated in the BIM-kit software toolkit. This approach is applied to 16 real-world passing lanes in Alberta, showcasing its utility in transportation. The analysis reveals significant enhancements, with rule compliance increasing from 61.82% to 91.31% after employing generative design techniques. These findings underscore the significance of generative design in transportation, offering engineers an efficient tool to create innovative, compliant solutions for highway projects. [ABSTRACT FROM AUTHOR]

Existing generative design cannot evaluate the perceived "aesthetic liking". Hence, quantification method of aesthetic liking, which can be modelled by the summation of "complexity" and "novelty", is needed. This study proposed the Gaussian curvature KL divergence as an index of novelty. This index represents the difference between a typical shape and a target shape on the basis of the free-energy principle. To confirm the effectiveness of the proposed index, we conducted cognitive experiments for shapes of an automobile generated by particle swarm optimization method with the proposed index as the objective function. The coefficient of determination between the proposed index and sensory evaluation values of "difference" were very high which support the applicability of the index ( R 2 = 0.86 ). By the summation of this index and a complexity index, a free-energy index is proposed, and its application for aesthetic liking is discussed. Finally, we confirmed that the Gaussian curvature KL divergence can be applied to evaluation of aesthetic liking. [ABSTRACT FROM AUTHOR]

Plain and simple, a corrugated cardboard material provides the right level of protection for goods in transit, which helps keep shipping costs down. As made mostly from wood material, a fairly renewable resource, corrugated cardboards are easily processed and also recycled into paper products and even new boxes. Moreover, thanks to the material's durability, many corrugated packages can be reused multiple times before hitting the recycle bin, which promotes eco-friendly corrugated packaging needs for the environment with no glues or extra binding chemical included inside the packaging at all. The research in this paper proposes a tentative design testbed for a product model of setter tray, wholly made in corrugated paperboard and seated inside packaging with no additional filling material, which, after all, aids realization of a totally eco-friendly corrugated packaging design. A methodical approach, a systematic functional design framework with design thinking methodologies such as function decomposition (FD), quality function deployment (QFD), and morphological matrix, has been introduced to generate a promising concept continuously refined onto a product model. The validation of mechanical properties of the designed product for its functions as desired has been conducted by using not only a computer-aided engineering (CAE) analysis of the computer-aided design (CAD) model, but also an engineering experiment of the product prototypes made of corrugated paperboard, followed by an in-depth discussion on design contributions and future works as well. [ABSTRACT FROM AUTHOR]

Topology-optimisation strategies for structural design of discrete surface structures result in unstructured patterns that require post-rationalisation to limit the number and complexity of the various structural elements. Fabrication-related objectives still demand further processing. Designers need topology exploration, before topology optimisation, following a generative-design approach that is decoupled from density tuning and shape design, to produce high-quality patterns. Such an approach allows to flexibly embed multiple design constraints, benefit from state-of-the-art form-finding algorithms and explore design trade-offs of the multiple requirements related to architecture, engineering and construction. This research investigates a parameterisation strategy to encode topological exploration of structured patterns based on quad meshes. The focus is set on singular vertices, which connect an irregular number of blocks, cables, or beams. Singularities are independent from pattern density and geometry, and have a fundamental influence on the qualitative and quantitative performance of the structure. We introduce an L-system encoding a quad-mesh grammar into a string that describes topological transformations of these singularities. Design applications to a net and a gridshell demonstrate the influence of singularities on the design of surface structures, and highlight the flexibility and generality of the approach in terms of geometrical processing and performance metrics. Beyond exploration, this parameterisation strategy opens to novel applications of search and optimisation methods for generative design of singularities in structural patterns. [ABSTRACT FROM AUTHOR]

Generative design (GD) techniques have been proposed to generate numerous designs at early design stages for ideation and exploration purposes. Previous research on GD using deep neural networks required tedious iterations between the neural network and design optimization, as well as post-processing to generate functional designs. Additionally, design constraints such as volume fraction could not be enforced. In this paper, a two-stage non-iterative formulation is proposed to overcome these limitations. In the first stage, a conditional generative adversarial network (cGAN) is utilized to control design parameters. In the second stage, topology optimization (TO) is embedded into cGAN (cGAN + TO) to ensure that desired functionality is achieved. Tests on different combinations of loss terms and different parameter settings within topology optimization demonstrated the diversity of generated designs. Further study showed that cGAN + TO can be extended to different load and boundary conditions by modifying these parameters in the second stage of training without having to retrain the first stage. Results demonstrate that GD can be realized efficiently and robustly by cGAN+TO. [ABSTRACT FROM AUTHOR]

As technology advances, architects often employ innovative, non-standard shapes in their designs for the fast-growing number of high-rise buildings. Conversely, climate change is bringing about an increasing number of dangerous wind events causing damage to buildings and their surroundings. These factors further complicate the already difficult field of structural wind analysis. Current methods for calculating structural wind response, such as the Eurocode, do not provide methods for unconventional building shapes or, in the case of physical wind tunnel test and in-depth computational fluid dynamics (CFD) simulation, they are prohibitively expensive and time-consuming. Thus, wind load analysis is often relegated to late in the design process. This paper presents the development of a computational method to analyze the effect of wind on the structural behavior of a 3D building model and optimize the external geometry to reduce those effects at an early design phase. It combines CFD, finite-element analysis (FEA), and an optimization algorithm in the popular parametric design tool, Grasshopper. This allows it to be used in an early design stage for performance-based design exploration in complement to the more traditional late-stage methods outlined above. After developing the method and testing the timeliness and precision of the CFD, and FEA portions on case study buildings, the tool was able to output an optimal geometry as well as a database of improved geometric options with their corresponding performance for the wind loading. [ABSTRACT FROM AUTHOR]

Campus master plans are released every few years for developing and implementing its physical infrastructure. Open spaces, compactness, connectivity, greenness, and environmental impact have often been the focus on its framework. In particular, the effect of new building development on existing buildings' occupant comfort and design intent is mostly ignored. Providing guidelines to retain existing users' comfort for stakeholders involved in design decision making will result in improved design decisions. Hence, this research aims to provide a work methodology to mitigate the adverse effects of new buildings on existing buildings' user comfort through a case study at Carleton University. The case study shows a methodology to retain the existing users' comfort by analyzing Carleton University's master plan on massing studies, occupant survey to understand their comfort needs, performance analysis of the impact of the new building on the existing building user comfort. The analysis reveals the key parameters to consider in design for occupants' comfort. Finally, the research reinforces the generative design and the need for dynamic modeling in campus master plans to mitigate the negative implications of new development on occupants' comfort. [ABSTRACT FROM AUTHOR]

Additive Manufacturing (AM) represents a new class of disruptive production and prototyping technologies. Topology optimization has been known to produce a high level of complexity in generated structure designs that is suitable for AM. Design for stress is a major factor in a number of scientific applications, where AM still experiences its own limitations that are unique to it. When such limitations and stress factors are not considered, postprocessing is generally required to modify generated designs such that it adds to cost and reduces performance of optimized structures. Hence, a stress-based optimization framework is proposed in conjunction to an AM model that incorporates the main trait of AM processes. With a density-based framework, the approach can be used with stress clustering to optimize the number of stress factors with a conventional topology optimization solver. The approach ensures the design of structures with restricted AM geometries that does not violate the filtering size constraints by use of stress clusters for compliance-based minimization problems. [ABSTRACT FROM AUTHOR]

Abstract Porous structures are useful elements in several applications, such as in tissue engineering and in the refinement of optimized elements, with the benefit of offering minimum weight, diminishing costs, and reducing material use. This paper presents a method to design porous structures in two dimensions based on grayscale images, in which the porosity and pore size can be changed as a function of a distribution of densities (gray intensity). Generative design tools, Voronoi diagrams and Delaunay triangulations, are developed in which the size of the cells is related to the gray intensity distribution within the image. The method was tested on grayscale images from X-ray and topological optimization algorithms to obtain CAD models suitable for additive manufacturing. These models were subsequently manufactured through an additive manufacturing process. The results show that the developed tool is able to successfully obtain cellular structures; moreover, the 3D-printed specimens were evaluated through both FEM and physical tests to compare which of the trabeculae shapes was better for loading transmission. [ABSTRACT FROM AUTHOR]

Purpose: This paper aims to propose a new framework on prioritizing and deployment of design for additive manufacturing (DfAM) strategies to an industrial component using Fuzzy TOPSIS multiple criteria decision-making (MCDM) techniques. The proposed framework is then applied to an automotive component, and the results are discussed and compared with existing design. Design/methodology/approach: Eight DfAM design alternatives associated with eight design criteria have been identified for framing new DfAM strategies. The prioritization order of the design alternatives is identified by Fuzzy TOPSIS MCDM technique through its closeness coefficient. Based on Fuzzy TOPSIS MCDM output, each of the design alternatives is applied sequentially to an automobile component as a case study. Redesign is carried out at each stage of DfAM implementation without affecting the functionality. Findings: On successful implementation of proposed framework to an automotive component, the mass is reduced by 43.84%, from 0.429 kg to 0.241 kg. The redesign is validated by finite element analysis, where von Mises stress is less than the yield stress of the material. Practical implications: The proposed DfAM framework and strategies will be useful to designers, R&D engineers, industrial practitioners, experts and consultants for implementing DfAM strategies on any industrial component without impacting its functionality. Originality/value: To the best of the authors' knowledge, the idea of prioritization and implementation of DfAM strategies to an automotive component is the original contribution. [ABSTRACT FROM AUTHOR]

In this article, we present a design methodology for resonant structures exhibiting particular dynamic responses by combining an eigenfrequency matching approach and a harmonic analysis-informed eigenmode identification strategy. This systematic design methodology, based on topology optimization, introduces a novel computationally efficient approach for 3D dynamic problems requiring antiresonances at specific target frequencies subject to specific harmonic loads. The optimization's objective function minimizes the error between target antiresonance frequencies and the actual structure's antiresonance eigenfrequencies, while the harmonic analysis-informed identification strategy compares harmonic displacement responses against eigenvectors using a modal assurance criterion, therefore ensuring an accurate recognition and selection of appropriate antiresonance eigenmodes used during the optimization process. At the same time, this method effectively prevents well-known problems in topology optimization of eigenfrequencies such as localized eigenmodes in low-density regions, eigenmodes switching order, and repeated eigenfrequencies. Additionally, our proposed localized eigenmode identification approach completely removes the spurious eigenmodes from the optimization problem by analyzing the eigenvectors' response in low-density regions compared to high-density regions. The topology optimization problem is formulated with a density-based parametrization and solved with a gradient-based sequential linear programming method, including material interpolation models and topological filters. Two case studies demonstrate that the proposed design methodology successfully generates antiresonances at the desired target frequency subject to different harmonic loads, design domain dimensions, mesh discretization, or material properties. [ABSTRACT FROM AUTHOR]

Without an explicit formulation to minimize support structures, topology optimization may create complex shapes that require an intensive use of support material when additively manufactured. We propose a neural network-based approach to topology optimization that aims to reduce the use of support structures in additive manufacturing. Our approach uses a network architecture that allows the simultaneous determination of an optimized: (1) part segmentation, (2) the topology of each part, and (3) the build direction of each part that collectively minimize the amount of support structure. Through training, the network learns a material density and segment classification in the continuous 3D space. Given a problem domain with prescribed load and displacement boundary conditions, the neural network takes as input 3D coordinates of the voxelized domain as training samples and outputs a continuous density field. Since the neural network for topology optimization learns the density distribution field, analytical solutions to the density gradient can be obtained from the input-output relationship of the neural network. We demonstrate our approach on several compliance minimization problems with volume fraction constraints, where support volume minimization is added as an additional criterion to the objective function. We show that simultaneous optimization of part segmentation along with the topology and print angle optimization further reduces the support structure, compared to a combined print angle and topology optimization without segmentation. [ABSTRACT FROM AUTHOR]

A generative design methodology is presented that solves for minimum volume and compliance space-frame systems, with consideration of stress and buckling constraints. The solution space is explored using formal topology optimisation routines. A parameterisation method converts voxelised topology optimisation solutions into skeletonised connectivity representations. An inequality constrained gradient descent optimisation method optimises and defines cross-sectional geometry. This enables fast and automatic solution generation, providing designers with sets of high-performing problem solutions. Skeleton representations provide an inexpensive modelling tool for parallel analysis of physical problems difficult to model using topology optimisation. Geometry is represented using traditional engineering cross-sections with well understood behaviour. This improves certainty in the performance of solutions, simplifying certification processes. The generative design of a structural aerospace bracket for additive manufacture has been used as a case study within this research. A design of experiments produced 360 topology optimisation results, altering input variables and discretisation resolution to identify their effect on solution outcomes and the performance of parameterisation. The proposed method was found to robustly generate high-performing solutions utilising low-resolution topology optimisation. Additionally, 6 high-performing topologies were identified, providing designers with a set of solutions to select from. Limitations on the parameterisation process were identified, with topology optimisation solutions with volume fractions above 0.2 not parameterising successfully, and simulations with a resolution of 5 mm element size and below generating low performing skeletonised topologies. [ABSTRACT FROM AUTHOR]

The density distribution in an urban matrix is one of the significant issues which affects other urban living factors such as building lighting, energy consumption and residents’ interactions. The research toward achieving the optimum density distribution has received attention for the last decade. However, developing a generative approach that provides more freedom for the formation of the plans and incorporates adaptability in different land blocks is still missing. To address such a gap, this study proposes an adaptable approach developing the formation of residential blocks. This formation is according to the pre-defined size and shape of the land, and sought performance objectives. Hence, a suite of applications including Grasshopper, Python and Ladybug were applied in a residential block of Tehran as a case study. The purpose is to develop a new density distribution increasing view quality, visual privacy, and solar gain. For the optimization process, a genetic algorithm was applied utilizing the topology optimization technique. The results of the optimization process highlight the significance of this research since the developed alternatives are more efficient in terms of improving the view quality, visual privacy and increasing the solar gain. This achievement expands the potential of this research to be applied in different case studies and with different design and development objectives in order to develop better shape plans of building blocks. [ABSTRACT FROM AUTHOR]

Since the advent of Industry 4.0, the manufacturing sector has had to face new challenges, which require the development of new skills and innovative tools. This scenario includes innovative production processes such as additive manufacturing (AM), a technology capable of producing a component layer-by-layer directly from the 3D model without needing specific tools during the building phase. Generative design (GD) may represent an opportunity to maximise the potential of AM techniques. GD is based on parametric computer-aided design (CAD) tools capable of generating multiple optimised outputs, among which the designer could select the most promising solution. This paper presents a general methodology for evaluating the GD outputs in the conceptual phase of design to select the best possible solution through a series of criteria at several levels. The evaluation method is deployed in an aerospace field case study. The procedure demonstrates the benefits of synergising GD with AM in the early stages of product development. This indicates that the developed methodology could reduce the number of iterations during the design process, and the result is a decrease in the overall time spent on the project, avoiding problems during the final stages of the design. [ABSTRACT FROM AUTHOR]

Biological systems in nature have evolved for millions of years to adapt and survive the environment. Many features they developed can be inspirational and beneficial for solving technical problems in modern industries. This leads to a specific form of design-by-analogy called bio-inspired design (BID). Although BID as a design method has been proven beneficial, the gap between biology and engineering continuously hinders designers from effectively applying the method. Therefore, we explore the recent advance of artificial intelligence (AI) for a data-driven approach to bridge the gap. This paper proposes a generative design approach based on the generative pre-trained language model (PLM) to automatically retrieve and map biological analogy and generate BID in the form of natural language. The latest generative pre-trained transformer, namely generative pre-trained transformer 3 (GPT-3), is used as the base PLM. Three types of design concept generators are identified and fine-tuned from the PLM according to the looseness of the problem space representation. Machine evaluators are also fine-tuned to assess the mapping relevancy between the domains within the generated BID concepts. The approach is evaluated and then employed in a real-world project of designing light-weighted flying cars during its conceptual design phase The results show our approach can generate BID concepts with good performance. [ABSTRACT FROM AUTHOR]

Generative design tools empowered by recent advancements in artificial intelligence (AI) offer the opportunity for human designers and design tools to collaborate in new, more advanced modes throughout various stages of the product design process to facilitate the creation of higher performing and more complex products. This paper explores how the use of these generative design tools may impact the design process, designer behavior, and overall outcomes. Six in-depth interviews were conducted with practicing and student designers from different disciplines who use commercial generative design tools, detailing the design processes they followed. From a grounded theory-based analysis of the interviews, a provisional process diagram for generative design and its uses in the early-stage design process is proposed. The early stages of defining tool inputs bring about a constraint-driven process in which designers focus on the abstraction of the design problem. Designers will iterate through the inputs to improve both quantitative and qualitative metrics. The learning through iteration allows designers to gain a thorough understanding of the design problem and solution space. This can bring about creative applications of generative design tools in early-stage design to provide guidance for traditionally designed products. [ABSTRACT FROM AUTHOR]

Generative adversarial networks (GANs) have shown remarkable success in various generative design tasks, from topology optimization to material design, and shape parametrization. However, most generative design approaches based on GANs lack evaluation mechanisms to ensure the generation of diverse samples. In addition, no GAN-based generative design model incorporates user sentiments in the loss function to generate samples with high desirability from the aggregate perspectives of users. Motivated by these knowledge gaps, this paper builds and validates a novel GAN-based generative design model with an offline design evaluation function to generate samples that are not only realistic but also diverse and desirable. A multimodal data-driven design evaluation (DDE) model is developed to guide the generative process by automatically predicting user sentiments for the generated samples based on large-scale user reviews of previous designs. This paper incorporates DDE into the StyleGAN structure, a state-of-the-art GAN model, to enable data-driven generative processes that are innovative and user-centered. The results of experiments conducted on a large dataset of footwear products demonstrate the effectiveness of the proposed DDE-GAN in generating high-quality, diverse, and desirable concepts. [ABSTRACT FROM AUTHOR]

Generative design is a system that automates part of the design process, but it cannot evaluate psychological issues related to shapes, such as "beauty" and "liking". Designers therefore evaluate and choose the generated shapes based on their experience. Among the design features, "complexity" is considered to influence "aesthetic preference". Although feature descriptors calculated from curvature can be used to quantify "complexity", the selection guidelines for curvature and feature descriptors have not been adequately discussed. Therefore, this study aimed to conduct a systematic classification of curvature and a feature descriptor of 3D shapes and to apply the results to the "complexity" quantification. First, we surveyed the literature on curvature and feature descriptors and conducted a systematic classification. To quantify "complexity", we used five curvatures (Gaussian curvature, mean curvature, Casorati curvature, shape index, and curvature index) and a feature descriptor (entropy of occurrence probability) obtained from the classification and compared them with the sensory evaluation values of "complexity". The results showed that the determination coefficient between the quantified and sensory evaluation values of "complexity" was highest when the mean curvature was used. In addition, the Casorati curvature tended to show the highest signal-to-noise ratio (i.e., a high determination coefficient irrespective of the parameters set in the entropy calculation). These results will foster the development of generative design of 3D shapes using psychological evaluation. [ABSTRACT FROM AUTHOR]

In a scaffold-based approach for bone tissue regeneration, the control over morphometry allows for balancing scaffold biomechanical performances. In this experimental work, trabecular geometry was obtained by a generative design process, and scaffolds were manufactured by vat photopolymerization with 60% (P60), 70% (P70) and 80% (P80) total porosity. The mechanical and biological performances of the produced scaffolds were investigated, and the results were correlated with morphometric parameters, aiming to investigate the influence of trabecular geometry on the elastic modulus, the ultimate compressive strength of scaffolds and MG-63 human osteosarcoma cell viability. The results showed that P60 trabecular geometry allows for matching the mechanical requirements of human mandibular trabecular bone. From the statistical analysis, a general trend can be inferred, suggesting strut thickness, the degree of anisotropy, connectivity density and specific surface as the main morphometric parameters influencing the biomechanical behavior of trabecular scaffolds, in the perspective of tissue engineering applications. [ABSTRACT FROM AUTHOR]

A streamlined design workflow that facilitates the efficient design and manufacture of patientspecific scaffolds independently applied by the surgical team has been recognised as a key step in a holistic approach towards the envisioned routine clinical translation of scaffold-guided bone regeneration (SGBR). A modular design workflow was developed to semi-automatically fill defect cavities, ensure patient specificity and ideal surgical scaffold insertion for a given surgical approach, add fixation points to secure the scaffolds to the host bone and generate scaffold based on Voronoi, periodic lattice and triply periodic minimal surface pore architectures. The adopted functional representation modelling technique produces models free from 3D printing mesh errors. It was applied to a clinical case of a complicated femoral bone defect. All models were free from mesh errors and the patient-specific fit and unobstructive insertion were validated via digital inspection and physical investigation by way of 3D printed prototypes. The real-time responsiveness of the workflow to user input allows the designer to receive real-time feedback from the surgeon, which is associated with reducing the time to finalise a patient-specific scaffold design. In summary, an efficient workflow was developed that substantially facilitates routine clinical implementation of SGBR through its ability to streamline the design of 3D printed scaffolds. [ABSTRACT FROM AUTHOR]

In the field of ocean engineering, the task of spatial hull modelling is one of the most complicated problems in ship design. This study presents a procedure applied as a generative approach to the design problems for the hull geometry of small vessels using elements of concurrent design with multi-criteria optimisation processes. Based upon widely available commercial software, an algorithm for the mathematical formulation of the boundary conditions, the data flow during processing and formulae for the optimisation processes are developed. As an example of the application of this novel approach, the results for the hull design of a sailing yacht are presented. [ABSTRACT FROM AUTHOR]

The idea of a walkable city refers to the extent to which the built environment encourages people to walk by establishing comfortable pedestrian routes, which allows people to connect to numerous services with reasonable effort and time. Walkability is currently regarded as a "good to know about" rather than a "must-have" factor for sustainable development. A combination of walkability with a standard design strategy, such as generative design, may result in a more efficient way of planning a walkable city. Interestingly, the sole indicator taken into account for walkability in the generative design domain is "distance to amenities", while in reality, other parameters, such as the comfort factor, could also influence walkability. Therefore, in this research, we developed a workflow based on the generative design, which considers the comfort dimension in combination with distance to amenities and street-level greeneries. We also included the human perspective, given that walkability is always personal. This research successfully generated three different scenarios of walkability-optimal urban plans, where the highest walkability is 82.43 (very walkable). Furthermore, the baseline scenario of two different locations also aligns with people's perspectives when compared. In addition, we found that the inclusion of a temporal dimension, enhanced 3D-related indicators, and constraints should benefit future research. [ABSTRACT FROM AUTHOR]

Modern day building design projects require multidisciplinary expertise from architects and engineers across various phases of the design (conceptual, preliminary, and detailed) and construction processes. The Architecture Engineering and Construction (AEC) community has recently shifted gears toward leveraging design optimization techniques to make well-informed decisions in the design of buildings. However, most of the building design optimization efforts are either multidisciplinary optimization confined to just a specific design phase (conceptual/preliminary/detailed) or single disciplinary optimization (structural/thermal/daylighting/energy) spanning across multiple phases. Complexity in changing the optimization setup as the design progresses through subsequent phases, interoperability issues between modeling and physics-based analysis tools used at later stages, and the lack of an appropriate level of design detail to get meaningful results from these sophisticated analysis tools are few challenges that limit multi-phase multidisciplinary design optimization (MDO) in the AEC field. This paper proposes a computational building design platform leveraging concurrent engineering techniques such as interactive problem structuring, simulation-based optimization using meta models for energy and daylighting (machine learning based) and tradespace visualization. The proposed multi-phase concurrent MDO framework is demonstrated by using it to design and optimize a sample office building for energy and daylighting objectives across multiple phases. Furthermore, limitations of the proposed framework and future avenues of research are listed. [ABSTRACT FROM AUTHOR]

This article discusses how generative design and 3D printing were used to redesign a payload part of a camera holder, with the goal of replacing the original polymer component with a metal printed component of greater strength. The article describes the investigation into the effect of different parameters of generative design on the results, including internal stress values and weight reduction, and the preparation of the parts for printing.

Owing to increasing attention and investigation of bio-inspired rules for generating innovative design with the aid of additive manufacturing, a concise introduction to bioinspired generative design is given. The state of the field is briefly reviewed, and open questions, challenges, and opportunities for further development are highlighted. [ABSTRACT FROM AUTHOR]

The growing performance of IT tools for the project allows to delegate entire phases of the conceptual and decision-making process thanks to the diffusion of algorithmic models referable to the category of artificial intelligence capable of automating important segments, if not the entire process of project production. The speed of evolution of these systems places us in a continuous transition phase in which it is necessary to reflect on the direction to be taken with respect to the many technical and IT opportunities. The text analyses the new instrumental and technical tools, reflecting on the limits and opportunities also with respect to the specific discipline of Architectural Technology, and on the need to reconsider the real meaning of their use. [ABSTRACT FROM AUTHOR]

This work presents a method for generating concept designs for coupled multiphysics problems by employing generative adversarial networks (GANs). Since the optimal designs of multiphysics problems often contain a combination of features that can be found in the single-physics solutions, we investigate the feasibility of learning the optimal design from the single-physics solutions, to produce concept designs for problems that are governed by a combination of these single physics. We employ GANs to produce optimal topologies similar to the results of level set topology optimization (LSTO) by finding a mapping between the sensitivity fields of specific boundary conditions, and the optimal topologies. To find this mapping, we perform image-to-image translation GAN training with a combination of structural, heat conduction, and a relatively smaller number of coupled structural and heat conduction data. We observe that the predicted topologies using GAN for coupled multiphysics problems are very similar to those generated by level set topology optimization, which can then be used as the concept designs for further detailed design. We show that using a combination of multiple single-physics data in the training improves the prediction of GAN for multiphysics problems. We provide several examples to demonstrate this. [ABSTRACT FROM AUTHOR]

The research activities and key results obtained at the artificial intelligence research center "Strong AI in the Industry" based on ITMO University are presented. The concept and goal setting of the Center, its key scientific results, open frameworks and libraries, and practical results demonstrating implementations in various industries are described. Special attention is given to activities concerning the development of promising AI technologies, such as automated machine learning, generative design, and planning of purposeful activity under uncertainty and data incompleteness. [ABSTRACT FROM AUTHOR]

The accuracy of meso-scaled thermomechanical method (TMM) in analysing distortion of hinge bracket additively manufactured using selective laser melting (SLM) is investigated with SS316L as material and optimised design using Generative Design (GD) approach in which loads, boundaries, and constraints are considered as parametric inputs and stress goal as output. Prior to 3D printing, process simulation for predicting distortion was conducted including a geometrical model with hexahedral mesh type, temperature dependent material and flow curve referred from specialised software. To reduce computational time, a lumping approach with volume fraction was implemented considering heat flux parameters. For validation purposes, a series of experiments was accomplished using SLM, manual support structure removal, and surface finishing process by sandblasting. Furthermore, the finished bracket was measured using an optical scanner and compared with the simulation result. In conclusion, TMM can predict the distortion accurately with percentage point-by-point and mean error of less than 3% and 1.5%, respectively. [ABSTRACT FROM AUTHOR]

In this paper, we present a predictive and generative design approach for supporting the conceptual design of product shapes in 3D meshes. We develop a target-embedding variational autoencoder (TEVAE) neural network architecture, which consists of two modules: (1) a training module with two encoders and one decoder (E2D network) and (2) an application module performing the generative design of new 3D shapes and the prediction of a 3D shape from its silhouette. We demonstrate the utility and effectiveness of the proposed approach in the design of 3D car body and mugs. The results show that our approach can generate a large number of novel 3D shapes and successfully predict a 3D shape based on a single silhouette sketch. The resulting 3D shapes are watertight polygon meshes with high-quality surface details, which have better visualization than voxels and point clouds, and are ready for downstream engineering evaluation (e.g., drag coefficient) and prototyping (e.g., 3D printing). [ABSTRACT FROM AUTHOR]

Offsite construction is gaining attention due to government policies promoting automation and productivity. Therefore, understanding the impact of building design on construction cost and the carbon footprint associated with offsite construction is important for the improved sustainability and climate resilience of the built environment. This study aims to develop a system approach, with the aid of building information modeling (BIM), for 3D geometric modeling and automatic generative design toward optimizing the carbon footprint and construction cost from offsite construction. A mathematical formulation is proposed to represent the topological relationships between different kinds of precast and cast-in-situ elements, which in turn, underpin 3D geometric modeling for possible geometric variations within precast buildings. New generative algorithms are developed to automatically manipulate building geometrics subject to pre-defined constraints, and create parametric BIM models in compliance with material types assigned by users. A BIM-based automation tool is developed to extract and match the model geometry/material information with a customized BIM object library, through which carbon emission factors and cost coefficients can be retrieved for multi-criteria sustainability analysis. The proposed new approach empowers 3D geometric modeling and geometry-based design automation, which enable comprehensive exploration of design alternates in precast construction. The proposed new method is illustrated via a case study that investigates the impact of different design variations on embodied carbon and construction cost of precast structures. The proposed BIM development takes around 30 minutes to create 1,000–1,500 new design options. Besides, it produces design options that contain a 30% less carbon footprint and construction cost than reference buildings from the literature. The results indicate that our proposed method can support automated design exploration at early stages, which help to identify optimal solutions for more informed decision making. [ABSTRACT FROM AUTHOR]

Machine Learning (ML) techniques have been used in an extensive range of applications in the field of structural and multidisciplinary optimization over the last few years. This paper presents a survey of this wide but disjointed literature on ML techniques in the structural and multidisciplinary optimization field. First, we discuss the challenges associated with conventional optimization and how Machine learning can address them. Then, we review the literature in the context of how ML can accelerate design synthesis and optimization. Some real-life engineering applications in structural design, material design, fluid mechanics, aerodynamics, heat transfer, and multidisciplinary design are summarized, and a brief list of widely used open-source codes as well as commercial packages are provided. Finally, the survey culminates with some concluding remarks and future research suggestions. For the sake of completeness, categories of ML problems, algorithms, and paradigms are presented in the Appendix. [ABSTRACT FROM AUTHOR]

Additive manufacturing processes have enabled the production of parts with complex geometry. In addition, novel design approaches such as generative design and crowdsourced design challenges enable the rapid generation of many feasible design alternatives with similar functionality but distinct geometry. In this study, we use an illustrative example, focused on laser-based powder bed fusion of metals, to explore how geometry and topology differences among parts with the same functionality can drive differences in cost. To accomplish this, we utilize a process-based cost model that can account for how variations in part geometry of different design alternatives impact the cost of the additive manufacturing process and associated post-processing operations. The cost model identified differences of up to 14% between the least and most expensive design alternatives. Part mass and build time were the most influential factors to group different design into relatively similar cost groups. High part complexity was associated with lower part cost, and was not strongly correlated to reject rates. Comparing designs within these groups showed several conflicting factors such as additive manufacturing and post-processing scrap and reject rates, which were geometry dependent. This result highlights the need for methods to better understand and quantify the effect of part geometry on manufacturing outcomes related to cost, including powder usage, post-processing requirements, and failure rates. Such methods can help designers to weigh tradeoffs between different cost, sustainability, quality, and performance objectives to select a preferred design alternative. [ABSTRACT FROM AUTHOR]

In recent years, the field of soft robotics has received considerable attention due to its potential in increasing the safety of human-robot interaction. The design of soft robots possesses great challenges. For example, the longstanding challenge of co-design morphology and actuation makes designing them by hand a trial-and-error process. Earlier work presented by the authors proposes a computational design synthesis (CDS) method for the automated design of virtual, soft locomotion robot morphologies. This work extends the CDS method for morphologies with the automated co-design of actuation. Two methods are considered. In the first method, the actuation of designs is described by parametric actuation curves (PACs) that model feedforward actuation patterns. For every morphology in the design process, a set of PACs is optimized that assumes symmetric and cyclic gaits. The second method, soft actor-critic (SAC) reinforcement learning, removes this assumption as well as models feedback control for comparison. Adding PAC optimization to the CDS method is shown to improve the performance of the resulting designs and to achieve better results within less design iterations. SAC is, however, deemed less effective, due to the need for design specific problem tuning for each new morphology. The SAC experiments also show that the best found soft robot gaits are symmetric and cyclic, although this is not a constraint in the SAC problem formulation, thus verifying the assumptions made in the PAC formulation. To validate the search space modeled in the co-design CDS method, a state-of-the-art soft robot is replicated and compared. [ABSTRACT FROM AUTHOR]

The integrated process of design and fabrication is invariably of particular interest and important to improve the quality and reduce the production cycle for structural joints, which are key components for connecting members and transferring loads in structural systems. In this work, using the generative design method, a pioneering idea was successfully realized to attain a reasonable configuration of the cross joints, which was then consecutively manufactured using 3D printing technology. Firstly, the initial model and generation conditions of a cross joint were constructed by the machine learning-based generative design algorithm, and hundreds of models were automatically generated. Then, based on the design objective and cost index of the cross joint, three representative joints were selected for further numerical analysis to verify the advantages of generative design. Finally, 3D printing was utilized to produce generative joints; the influences of printing parameters on the quality of 3D printing are further discussed in this paper. The results show that the cross joints from the generative design method have varied and innovative configurations and the best static behaviors. 3D printing technology can enhance the accuracy of cross joint fabrication. It is viable to utilize the integrated process of generative design and 3D printing to design and manufacture cross joints. [ABSTRACT FROM AUTHOR]

Automated design synthesis has the potential to revolutionize the modern engineering design process and improve access to highly optimized and customized products across countless industries. Successfully adapting generative machine learning to design engineering may enable such automated design synthesis and is a research subject of great importance. We present a review and analysis of deep generative machine learning models in engineering design. Deep generative models (DGMs) typically leverage deep networks to learn from an input dataset and synthesize new designs. Recently, DGMs such as feedforward neural networks (NNs), generative adversarial networks (GANs), variational autoencoders (VAEs), and certain deep reinforcement learning (DRL) frameworks have shown promising results in design applications like structural optimization, materials design, and shape synthesis. The prevalence of DGMs in engineering design has skyrocketed since 2016. Anticipating the continued growth, we conduct a review of recent advances to benefit researchers interested in DGMs for design. We structure our review as an exposition of the algorithms, datasets, representation methods, and applications commonly used in the current literature. In particular, we discuss key works that have introduced new techniques and methods in DGMs, successfully applied DGMs to a design-related domain, or directly supported the development of DGMs through datasets or auxiliary methods. We further identify key challenges and limitations currently seen in DGMs across design fields, such as design creativity, handling constraints and objectives, and modeling both form and functional performance simultaneously. In our discussion, we identify possible solution pathways as key areas on which to target the future work. [ABSTRACT FROM AUTHOR]

Tire tread patterns have played an important role in the automotive industry because they directly affect automobile performances. The conventional tread pattern development process has successfully produced and manufactured many tire tread patterns. However, a conceptual design process, which is a major part of the whole process, is still time-consuming due to repetitive manual interaction works between designers and engineers. In the worst case, the whole design process must be performed again from the beginning to obtain the required results. In this study, a deep generative tread pattern design framework is proposed to automatically generate various tread patterns satisfying the target tire performances in the conceptual design process. The main concept of the proposed method is that desired tread patterns are obtained through optimization based on integrated functions, which combine generative models and tire performance evaluation functions. To strengthen the effectiveness of the proposed framework, suitable image pre-processing, generative adversarial networks (GANs), two-dimensional (2D) image-based tire performance evaluation functions, design generation, design exploration, and image post-processing methods are proposed with the help of domain knowledge of the tread pattern. The numerical results show that the proposed automatic design framework successfully creates various tread patterns satisfying the target tire performances such as summer, winter, or all-season patterns. [ABSTRACT FROM AUTHOR]

For developing performing medical devices, the biomechanical adaptation plays a key role for both, the bulk material (at macro scale) and surfaces (at micro/nano scale). As concerning the interaction of living cells with artificial surfaces, the cellular mechano‐sensitivity has direct effects on tissue structure and functions. In this work, surfaces architectures, obtained by self‐nano‐crystallization, were realized on a gum‐alloy as substrate (titanium‐niobium‐zirconium‐iron‐oxygen), using silicone masks with micrometric patterns/perforations, on which a near‐surface‐severe‐plastic‐deformation process was applied. For finding 3D suitable geometries of the perforations and their distribution on the masks, generative design algorithms were developed using dedicated software tools. The masks were obtained by 3D printing using the fused filament fabrication method. Multidirectional repeated mechanical impacts on the sample surface with high velocity balls at high strain rates were applied. Deeper compressive residual stresses are generated, creating a superficial self‐nano‐crystallized structure on un‐covered areas, due to small volume‐multidirectional local plastic deformation. The applied parameters have a direct effect on local micro‐topography and microstructure evolution. [ABSTRACT FROM AUTHOR]

In this article, we propose an equivalent circuit theory-assisted deep learning approach to accelerate the design of metasurfaces. By combining the filter equivalent circuit theory and a sophisticated deep learning model, designers can achieve efficient metasurface designs. Compared with most existing metasurface generative design methods that rely on arbitrarily generated training dataset (TDS), the proposed method can adaptively produce highly relevant and low-noise training samples under the guidance of filter equivalent circuit theory, resulting in a significantly narrowed target solution space and improved model training efficiency. Furthermore, we select the variational autoencoder (VAE) as a generative model, which can compress the raw training samples into a lower-dimensional latent space where optimization methods, such as genetic algorithm, can be more efficiently executed to find the optimal design than a brute-force search. To verify the effectiveness of the proposed method, we apply it in the creation of three examples of frequency selective surfaces (FSSs), presenting wide-band, dual-band, and band-stop responses. Experimental results show that the proposed method can realize much faster and more stable convergence than deep learning design methods without domain knowledge. [ABSTRACT FROM AUTHOR]

Although additive manufacturing can produce nearly any geometry, users have limited choices in the designs. Topology optimization can create complex shapes, but it provides only one solution for one problem, and existing design exploration methods are ineffective when the design space is huge and high-dimensional. Therefore, this paper develops a new generative design method to improve the diversity of topology-optimized designs. Based on the observation that topology optimization places materials along the principal directions to maximize stiffness, this paper creates a rule of principal direction and applies it to swarm intelligence for form-finding. The shapes got by the swarming process possess both randomness and optimality. After they are further optimized, the final designs have high diversity. This is the first time integrating structural stiffness as a swarm principle to influence the collective behavior of decentralized, self-organized systems. The experimental results show that this method can generate interesting designs that have not been seen in the literature. Some results are even better than those got by the original topology optimization method, especially when the problem is more complex. This work not only allows users to choose unique designs according to their preference, but also helps users find better designs for their application. [ABSTRACT FROM AUTHOR]

Recent advances in generative modeling, namely Diffusion models, have revolutionized generative modeling, enabling high-quality image generation tailored to user needs. This paper proposes a framework for the generative design of structural components. Specifically, we employ a Latent Diffusion model to generate potential designs of a component that can satisfy a set of problem-specific loading conditions. One of the distinct advantages our approach offers over other generative approaches is the editing of existing designs. We train our model using a dataset of geometries obtained from structural topology optimization utilizing the SIMP algorithm. Consequently, our framework generates inherently near-optimal designs. Our work presents quantitative results that support the structural performance of the generated designs and the variability in potential candidate designs. Furthermore, we provide evidence of the scalability of our framework by operating over voxel domains with resolutions varying from 3 2 3 to 12 8 3 . Our framework can be used as a starting point for generating novel near-optimal designs similar to topology-optimized designs. • Latent diffusion model for generating 3D structural component designs. • Framework for generating component designs consistent with topology optimization. • Generated designs have similar (near-optimal) strain energy to SIMP designs. • Large scale 3D voxel dataset for structural topology optimization. [ABSTRACT FROM AUTHOR]