Authors
Jorge Luis Manzoni, Expert Sector Consultant | Infrastructure, Services & Real Estate at NTT DATA | LinkedIn profile
Jesús de Paz, Director of Infrastructure, BIM & Digital Transformation Department at INGECID | LinkedIn profile
Pablo Nistal, Head of Generative Design Area at INGECID | LinkedIn profile
(Note: Opinions in the articles are of the authors only and do not necessarily reflect the opinion of the EU)
The journey of architectural and engineering design has undergone a remarkable transformation, evolving from traditional methods to the sophisticated realms of parametric and generative design powered by Artificial Intelligence (AI). In the past, traditional design relied heavily on manual drafting techniques, in which architects and engineers produced 2D representations of buildings and infrastructures without integrating additional data or information. Although foundational, this approach often limited the ability to visualise complex structures and effectively optimise designs.
The advent of Computer-Aided Design (CAD) marked a significant shift, introducing 2D and 3D modelling capabilities that enabled more detailed project representations. However, these models often lacked the intelligence and integrated data necessary for comprehensive analysis and effective decision-making. The emergence of parametric design transformed the landscape by allowing models to adapt dynamically to parameter changes, promoting a more responsive and flexible design process. Building Information Modelling (BIM) advanced this evolution by integrating data and processes and enhancing collaboration among stakeholders, thereby streamlining workflows and improving overall project outcomes.
As we look towards the near future, AI-powered generative design emerges as the next frontier in BIM. This innovative approach leverages advanced algorithms and AI to automatically generate design options based on defined criteria. By incorporating controlled automatic option generation, design optimisation, alternative selection, and positive reinforcement learning, generative design enables architects and engineers to explore numerous possibilities, ultimately delivering more efficient, sustainable, and aesthetically appealing solutions. This evolution not only enhances creativity but also revolutionises the way we conceive and construct our built environment.
In this context, the European Union's regulatory framework for AI, as outlined in the EU Digital Strategy, plays a crucial role in ensuring the responsible development and deployment of AI technologies. Moreover, the emphasis on EU competitiveness highlights the importance of fostering innovation in the construction sector to keep Europe at the forefront of technological advancement. Aligning generative design with these policy initiatives will establish a robust foundation for a more sustainable and competitive built environment.
In today’s architectural and engineering landscape, parametric design has emerged as a powerful tool for streamlining the configuration of buildings and infrastructure. This approach allows designers to create models that are not only visually compelling but also highly adaptable to evolving requirements. At the heart of parametric design lies the concept of object libraries—predefined components that automatically adjust based on input data. This flexibility enables architects and engineers to modify parameters such as dimensions, materials, and performance criteria, resulting in a more efficient design process.
One of the key advantages of parametric design is its capacity to support rapid iteration. By adjusting parameters, designers can immediately visualise the impact of changes on the overall project, encouraging a more exploratory and creative design approach. This capability is particularly beneficial in complex projects that must consider multiple variables, including site conditions, regulatory requirements, and client preferences.
To further enhance the capabilities of parametric design, many software platforms now support API code development. This functionality allows users to create custom scripts and automate repetitive tasks, significantly improving productivity. For example, designers can develop algorithms that generate multiple design options based on defined criteria, enabling them to explore a broader range of possibilities without the manual effort typically required.
Additionally, Parametric Family Design plays a vital role in this process. By creating families of components that can be easily modified, designers can maintain consistency across projects while retaining the flexibility to adapt to specific challenges. This approach not only saves time but also enhances collaboration, as all team members can work with a shared set of adaptable components.
Visual programming is another powerful tool that complements parametric design. Using graphical interfaces to represent design logic, architects and engineers can intuitively connect different elements of their models and manipulate parameters based on data, measurements, or external inputs. This method democratises the design process, enabling individuals with varying levels of technical expertise to contribute meaningfully.
However, parametric design requires architects and engineers to work actively, providing targeted inputs and making decisions throughout the design process. By leveraging the efficiency of parametric tools, they can test multiple alternatives until they find those that meet their requirements and needs. To take it a step further, it is necessary to make the leap to generative design.
Generative design represents a paradigm shift in the architectural and engineering fields, harnessing the power of advanced algorithms and AI to create innovative solutions. At its core, generative design works by utilising AI, data sources, and design variables, allowing designers to input specific criteria and constraints that guide the process. This data may include a wide range of factors, such as site conditions, material properties, environmental considerations, regulatory requirements, and performance metrics—all of which play a crucial role in shaping the final outcome.
Once the input data and potential design variables are defined, the next step is to translate the overall project concept into a coherent parametric model. This goes beyond modelling individual components and requires defining relationships and dependencies between different elements of the design. It involves a level of abstraction that enables the model to reflect the project’s structure and constraints in a manageable and adaptable way. This foundational step is essential for allowing the generative design process to explore alternative solutions that are both creative and aligned with the project’s goals and requirements.
The next step involves developing a generative algorithm capable of producing design alternatives by varying parameters within the established relationships. Rather than simply generating a fixed set of options, the algorithm iteratively explores the design space: it creates a series of solutions, evaluates them against predefined performance criteria, and adjusts the parameters based on the outcomes. This feedback-driven process continues over multiple generations, with the goal of progressively improving solutions according to specific optimisation objectives—whether related to performance, aesthetics, cost, or sustainability.
In this context, the algorithm functions as both a generator and an evaluator, balancing creativity with performance. Advanced analytical tools are integrated into this loop to assess each iteration, enabling designers to identify and converge on high-performing solutions that may be difficult to achieve through conventional methods.
The benefits of generative design are manifold. It enhances creativity by enabling designers to explore a wider range of possibilities, often resulting in innovative solutions that challenge conventional design boundaries. Moreover, it significantly reduces the time spent on manual iterations, allowing teams to focus on higher-level decision-making and strategic planning. Additionally, the data-driven nature of generative design fosters collaboration among stakeholders, as all parties can engage with a shared set of insights and analyses.
Figure 1. Algorithmic Iteration.
To illustrate the application of generative design, consider the structural design of a data centre (DC), where cost and efficiency are critical. Beginning with inputs such as required floor space, functional zones, and load conditions, the algorithm explores design options by adjusting variables including the number of spans, beam lengths, number of floors, and types of structural sections.
Each alternative is evaluated based on cost, material use, and its impact on operational efficiency—such as equipment accessibility or layout flexibility. Through iterative optimisation, the process converges on solutions that meet all technical requirements while minimising structural costs and maximising the performance of the DC.
Optimising space distribution within a building is another compelling example of generative design. Imagine a scenario in which an architect aims to design a multi-functional space that accommodates various activities while maximising natural light and minimising energy consumption. By inputting parameters such as room dimensions, desired functions, location, and environmental factors (such as sunlight exposure and airflow), the generative design algorithm can explore countless configurations for room layouts and orientations.
The algorithm evaluates each design against performance metrics such as spatial efficiency, user experience, and energy efficiency. This process enables the architect to identify the most effective arrangement of spaces—one that not only meets the functional requirements but also promotes sustainability through optimal resource use and energy conservation.
In both examples, generative design demonstrates its ability to analyse complex relationships between input data and design variables, ultimately leading to innovative solutions that traditional methods may overlook. By leveraging data and advanced algorithms, generative design empowers architects and engineers to create structures that are not only functional and aesthetically appealing but also optimised for performance and sustainability.
Tools such as the one developed by INGECID—built on their proprietary Common Data Environment (CDE) platform, Vircore—are unlocking the potential of generative AI to empower designers and architects to create more efficient solutions for both new and existing buildings, while reducing costs. This innovative technology enables the exploration of a wide range of design options, optimising performance and sustainability in architectural and engineering process.
Generative design has a wide range of applications across various fields, demonstrating its versatility and effectiveness in optimising complex design challenges. Below are some notable use cases in building design:
In conclusion, the integration of AI in generative design marks a revolutionary advancement in the fields of architecture and engineering. By harnessing AI's computational power, designers can explore an infinite array of solutions that optimise performance, sustainability, and aesthetics. This technology not only enhances creativity but also streamlines the design process, enabling rapid iterations and data-driven decision-making. As generative design continues to evolve, its reliance on AI will empower professionals to tackle increasingly complex challenges, ultimately resulting in innovative structures that meet the demands of a sustainable future.
Moreover, the European Union's support for AI—through initiatives like the AI Pact—highlights its commitment to fostering innovation and competitiveness in the sector. This alignment of policy and technological progress underscores the importance of generative design in shaping a sustainable built environment, ensuring that Europe remains a leader in integrating AI into architecture and engineering. The synergy between AI and generative design is set to redefine the built environment while contributing to the EU's broader goals of competitiveness and sustainability.
