Generative Design
Intro
What is generative design, and how is it influencing architectural design? This research traces the origins of generative design thinking and examines its growing impact on design workflows within the AEC industry. We explore how the integration of computational logic and generative methodologies can help us to address complex design challenges. The research also identifies the current limitations of these systems and illustrates how we approach generative design in practice at Park, showcasing two real-world case studies that highlight both the potential and the challenges of applying these methodologies to architectural projects.
Defining Generative Design
Generative design is a computational design methodology in which outcomes emerge through the application of rule-based algorithmic frameworks. Designers set up the initial parameters, such as geometrical relationships or performance targets, and define algorithmic logics that allow them to iteratively explore multiple design solutions. Rather than crafting each solution directly, this methodology allows us to translate an initial idea into a structured sequence of operations, creating a dynamic system that autonomously generates multiples design alternatives. By embedding evaluation criteria into the generative logic and leveraging specialized algorithms (i.e. evolutionary algorithms), designs can be tested and optimized against predefined objectives, leading to solutions that might not have been conceived through conventional design approaches.
Generative Design Toolkit
The toolkits available to architects have evolved alongside technological advancements. Until the 1980s, design was a manual process with no digital mediation. The introduction of Computer-Aided Design (CAD) in the 1980s marked a first significant shift, enabling designers to draft and store drawings in a digital environment. In the early 2000s, Algorithm-Aided Design (AAD) further evolved this process, moving from manual design manipulation to rule-based design processes. Designers could define parameters and constraints, enabling parametric models to generate design variations. Recently, generative design has gained traction in the AEC industry, thanks also to the integration of generative tools into the most common 3D-modeling software. The increased accessibility to these tools has allowed designers to explore iterative workflows and incorporate evolutionary algorithms, like genetic algorithms (GA), to develop performance-driven design solutions.
Emerging Design paradigms
Generative design introduces a shift in architectural practice, enabling designers to move beyond traditional workflows by incorporating computational methods that optimize design performance and help addressing complex design challenges. Three key approaches illustrate its integration into the design process:
Form-driven design:
utilizes computational tools to generate and manipulate complex geometries, expanding possibilities for spatial articulation and aesthetic expression.
Performance-driven design:
focuses on optimizing measurable criteria such as energy efficiency, structural integrity, and environmental performance.
Nature-driven design:
inspired by biomimicry, translates biological principles into architectural strategies, promoting the adoption of systems and solutions inspired by natural processes.
Whereas each approach prioritizes different aspects of the design intent, they do not exclude one another. Understanding their differences allows architects to strategically apply computational and generative tools, balancing creative exploration with functional optimization in the design process.
Our approach
At Park, our approach to computational and generative design transcends the boundaries of formalism, positioning these methodologies as essential tools to amplify design intent and enhance project outcomes. For us, these tools are not just meant to create and manage complex geometries but strategic instruments to address specific design challenges, improve the design process efficiency and unlock its creative potential. The research presents two case studies illustrating how we at Park have leveraged generative design methodologies to address diverse architectural challenges. These examples not only illustrate the potential of these tools to unlock innovative solutions but also the challenges we navigated in developing tailored workflows to suit each project’s unique demands. The first case study illustrates how the use of generative design tools enabled us to conduct in-depth evaluations of how a building’s massing interacts with its context while the second focuses on the application of a generative design process to the development of a building façade, tracing the integration of these tools from the earliest concept design stages.
Key challenges and opportunities
The two case studies highlight both the potential and the challenges of implementing generative design principles in real-world projects. A major hurdle is the resource-intensive nature of this approach, requiring significant time and computational power, which can impact tight project schedules. Current tools are also limited by a lack of holistic models, focusing on specific, quantifiable parameters, making it difficult to address multiple design challenges simultaneously. The iterative process adds further complexity, as each design change triggers new solution cycles. To optimize this process and maintain coherence throughout project development, it’s crucial to define clear design objectives and embed relevant parameters from the outset. This ensures that generative models align with the original design intent while addressing practical constraints, such as constructability. When carefully integrated, generative design tools enhance design efficiency, support data-driven insights, and enable the exploration of possibilities that traditional methods might overlook.
Research by Giulio Dini, Daniele Ferrari