Abstract:The conceptual design decisions have the largest influence on a building project's safety, value, and environmental impact; hence they are commonly assigned to a "senior engineer" to make use of his/her experience. However, the senior engineers can be biased towards solutions inside their area of expertise, which often prevents them from finding the best solutions among alternatives that must consider complex inter-related, and multi-disciplinary parameters. The engineering community could benefit from a rapid and high-quality decision-making method or tool to increase the speed and quality of its high-impact design choices. There are valuable studies in the literature exploiting Artificial Intelligence (AI) to improve the structural design process; however, most of them focus on the final design stage (e.g., Building Information Modeling), and the rest requires an existing project database (e.g., architectural drawings, already decided material types) to propose a small number of initial design alternatives. In this article, we present the development and validation of a genetic algorithm tool based on Non-dominated Sorted Genetic Algorithm II (NSGA-II) that can be used to analyse a wide range of safe, economical and low-CO2 options for the conceptual design of buildings. The design space starts from a design brief (with only the information about the site characteristics and project objectives). The solutions are explored with the material, grid size, floor type, lateral resistance, and foundation system variables. In a short computational time (< 2 min per run), users are provided with a Pareto graph of a large set of feasible solutions (in terms of cost, embodied CO2 emissions and free space) that an engineer would not be typically able to evaluate within a traditional conceptual design process. For future applications, the methodology presented in this paper is flexible to include more engineering materials (e.g., timber, masonry, structural glass), complex architectural forms and merge other disciplines in decision making (e.g., building physics construction management, fire safety).

A hybrid machine learning and evolutionary approach to material selection and design optimization for eco-friendly structures

Integrating Material Selection with Design Optimization via Neural Networks

Material Selection Combined with Optimal Structural Design for Mechanical Parts

Machine learning-enabled constrained multi-objective design of architected materials

Designing Bioinspired Composite Structures via Genetic Algorithm and Conditional Variational Autoencoder

A semi-supervised approach to architected materials design using graph neural networks

A survey of machine learning techniques in structural and multidisciplinary optimization

Learning to simulate and design for structural engineering

Machine learning-assisted design of material properties

Accessing Complex Reconstructed Material Structures with Hybrid Global Optimization Accelerated via on-the-fly Machine Learning

An efficient optimization approach for designing machine learning models based on genetic algorithm

Machine learning-enabled design of architected materials

Material Microstructure Design Using VAE-Regression with Multimodal Prior

Design Space Exploration and Explanation via Conditional Variational Autoencoders in Meta-Model-Based Conceptual Design of Pedestrian Bridges

A genetic algorithm tool for conceptual structural design with cost and embodied carbon optimization

Machine Learning Approaches in Agile Manufacturing with Recycled Materials for Sustainability

Accelerated Discovery of Sustainable Building Materials

Material Performance Prediction and Mechanical Structure Optimization Design Algorithm Based on Multi-Source Heterogeneous Data Fusion

Low-carbon design of structural components by integrating material and structural optimization

Data-driven topology optimization using a multitask conditional variational autoencoder with persistent homology

Machine Learning-Based Multi-Objective Design Exploration Of Flexible Disc Elements