To address the increased energy demands and carbon emissions caused by global urbanization, it is imperative to seek high-performance urban design solutions. Urban form generation and optimization ...(UFGO) is a powerful way of supporting performance-driven urban design by strategically searching for a possible design space to approach optimal solutions. Relevant areas of urban form generative design, urban energy and environment simulation, and urban form optimization have been widely studied. However, UFGO, which integrates these parts into an effective workflow, is still an emerging and meaningful research field lacking a systematic review. We examined studies that utilized UFGO techniques for urban design at different scales and outlined the general workflow. An overview of the available methods and tools, as well as their basic principles for each step, namely, urban form generation, performance simulation, and optimization, is provided. The reader will be well versed in the key problems and technical paths of UFGO. According to the review, UFGO is technically feasible; nevertheless, existing limitations necessitate further exploration. Future studies should focus on developing user-friendly UFGO software packages for urban designers, systematic and flexible generative design methods, and efficient data-driven models for urban performance evaluation. In addition, an evaluation system for UFGO techniques is also required to facilitate comparative studies and the widespread application of UFGO techniques.
•Studies that applied or discussed UFGO techniques integrally are reviewed.•A typical UFGO workflow, as well as methods and tools for each step are summarized.•The efficiency and accuracy of urban performance simulation methods are discussed.•Current technical paths and key issues for performance-driven UFGO are summarized.
Textured surface is of fundamental and practical importance in numerous emerging applications due to its beneficial effects on the tribological performance. In this work, a machine learning based ...universal generative design framework is proposed for surface texturing designing by combining specific convolutional neural network with improved Monte Carlo search. The optimal patterns of surface texture generated by machine learning are systematically studied under different conditions. Our results show that the machine generated wavy and chevron-like textures have the potential to dramatically improve the tribological performance of sliding surface with infinite design domain. Compared with the reported optimal texture, the friction coefficient of machine generated texture is reduced to 27.3∼49.7%, and the load carrying capacity is increased to 126.1∼144.4%.
Display omitted
•Machine learning based method is efficient for optimizing surface texture.•BN free CNN is employed to predict tribological performance of surface texture.•Monte Carlo search is improved to efficiently find optimal surface texture.•Optimal surface texture is generated by machine learning without human experience.•Machine generated textures could dramatically improve tribological performance.
A modular housing design entails an expensive and time-consuming process including iterative modification steps for satisfying various project and modular construction requirements. In addition, the ...fulfillment of all functional requirements within a limited budget in the modular housing design process remains elusive. A lack of a systematic approach for module configuration is an another critical obstacle making the design procedure more arduous and complicated. To ameliorate these knowledge and practice gaps, this study provides a new coupled generative adversarial network (CoGAN)-based framework for automated modular housing design generation. Furthermore, this approach encompasses a new module configuration algorithm that structurally modularize a generated housing design layout. This framework is expected to contribute to establishing the body of knowledge in a generative design of modular housing for mass building production and help architects and relevant stakeholders facilitate their design processes by yielding feasible, constructible, and optimal modular housing design alternatives.
•A new deep learning-based method to automate modular housing design considering constraints of manufacturing and assembly•A flexible automated module configuration approach compatible with the current practices in the volumetric modular housing•Creation of 3D building information models of the generated housing designs to achieve higher design-production integration.•Prototype development and validation of benefits by executing the integrated generative modular housing design system•Establish a fundamental framework for developing context-based design generations using the AI technique.
Existing systems that employ Automatic Speech Recognition (ASR) technology to retrieve information from the BIM model fail to provide remote interaction, retrieve a wide range of data, and automate ...the entire process. This is particularly a problem for users with disabilities. The paper offers a two-way, automated, and agnostic solution to this theoretical and methodological gap. A ‘Proof of Concept’ prototype was developed using Amazon Alexa – as the AI voice assistant platform – to test the applicability. The outcome shows that the created and the retrieved information is valid. Furthermore, there is a high level of interoperability among the components of the proposed solution, including the AI voice assistant interface and mediation environment to convert verbal requests and retrieve information to CSV files. Future research will extend the created solution to retrieve and access information from a BIM cloud model.
•An automated data retrieval-based AI voice assistant for BIM users is provided.•Interacting with the BIM model remotely via a voice assistant interface has been enabled.•Practitioners with vision disabilities can receive and add information to BIM models.•Novice BIM users can practice BIM features and retrieve information with minimal experience level.
In many reinforcement learning tasks, the goal is to learn a policy to manipulate an agent, whose design is fixed, to maximize some notion of cumulative reward. The design of the agent's physical ...structure is rarely optimized for the task at hand. In this work, we explore the possibility of learning a version of the agent's design that is better suited for its task, jointly with the policy. We propose an alteration to the popular OpenAI Gym framework, where we parameterize parts of an environment, and allow an agent to jointly learn to modify these environment parameters along with its policy. We demonstrate that an agent can learn a better structure of its body that is not only better suited for the task, but also facilitates policy learning. Joint learning of policy and structure may even uncover design principles that are useful for assisted-design applications.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
•A novel generative design approach is proposed for affordable urban cool spots.•Random Forest is deployed to accelerate simulating outdoor thermal environment.•Principal Component Analysis merges ...cost indicators into one objective function.•Genetic algorithm generates mass-based design layouts based on the Pareto front.•Average temperature decreases by 0.17 °C and costs by 82.57% in the case study.
Whilst designing cool small neighborhoods (called ‘cool spots’ in this paper) remains an enormous technical challenge, clients and their designers are also confronting with the perpetual burden of the financial sphere. This research aims to develop a novel methodological approach for designers to search for affordable cool spots in dense urban areas. It does so by conducting genetic combinatorial optimizations augmented by Random Forest (RF) and Principal Component Analysis (PCA) algorithms. What is particularly innovative is to develop a mass-based generative design approach to produce neighborhood options for the subsequent combinatorial optimization. The methodology is tested in a real-world urban renewal project in Hong Kong, which is epitomized by high density and hot and humid weather in the summer. The results show that the design approach can automatically identify high-performance schemes of cool spot design, reducing the daily average thermophysiological equivalent temperature from averagely 29.76 °C to at lowest 29.59 °C, and decreasing the construction cost by 82.57%. With proper translation, the approach can serve as a useful and robust design assisting tool for designing and developing cool and cost-aware buildings and neighborhoods in urban areas.
•The structures and nanomechanical properties of the veins provide inspiration for bionic engineering designs.•This paper presents the simulation analysis of the structural statics and aerodynamics ...of 3D coupling model (HW-I and HW-II).•The 3D bionic wing (BioW) model was established and optimized using the generative design method.
In view of the application prospect of the hindwing of Anomala Corpulenta Motschulsky in the field of foldable Micro Aerial Vehicles (MAVs), this paper investigated the morphology, macro/microstructure of the hindwing, and the nanomechanical properties of the wing veins and the wing membrane. It revealed the variation of nanohardness and elastic modulus between different veins and different positions of the same wing veins. This paper established a 3D coupling model of the hindwing based on the principle of coupling bionics. This paper presents a simulation analysis of the structural statics (uniform load distribution) and aerodynamics (under different attack angles, flight velocities, and flapping frequencies). Two 3D coupling models (HW-I and HW-II) of the hindwing were discussed the deformation and flight aerodynamic performance of Workbenches and Fluent. On that basis, the bionic wing was generatively designed, and a 3D bionic wing (BioW) model was established using the generative design method. Simulation analyses were performed through structural statics and aerodynamics. The results showed that the stress distribution was relatively uniform and that the overall displacement deformation was minimal for the BioW model. Moreover, the BioW model had better flight efficiency and aerodynamic performance.
This work proposes a novel generative design tool for passive grippers---robot end effectors that have no additional actuation and instead leverage the existing degrees of freedom in a robotic arm to ...perform grasping tasks. Passive grippers are used because they offer interesting trade-offs between cost and capabilities. However, existing designs are limited in the types of shapes that can be grasped. This work proposes to use rapid-manufacturing and design optimization to expand the space of shapes that can be passively grasped. Our novel generative design algorithm takes in an object and its positioning with respect to a robotic arm and generates a 3D printable passive gripper that can stably pick the object up. To achieve this, we address the key challenge of jointly optimizing the shape and the insert trajectory to ensure a passively stable grasp. We evaluate our method on a testing suite of 22 objects (23 experiments), all of which were evaluated with physical experiments to bridge the virtual-to-real gap. Code and data are at https://homes.cs.washington.edu/~milink/passive-gripper/
Engineering design research integrating artificial intelligence (AI) into computer-aided design (CAD) and computer-aided engineering (CAE) is actively being conducted. This study proposes a deep ...learning-based CAD/CAE framework in the conceptual design phase that automatically generates 3D CAD designs and evaluates their engineering performance. The proposed framework comprises seven stages: (1) 2D generative design, (2) dimensionality reduction, (3) design of experiment in latent space, (4) CAD automation, (5) CAE automation, (6) transfer learning, and (7) visualization and analysis. The proposed framework is demonstrated through a road wheel design case study and indicates that AI can be practically incorporated into an end-use product design project. Engineers and industrial designers can jointly review a large number of generated 3D CAD models by using this framework along with the engineering performance results estimated by AI and find conceptual design candidates for the subsequent detailed design stage.