In building design, the structural design is usually incorporated in the latter stages of the design, and the computational tools currently available focus more on converging and optimizing a single solution rather than providing an opportunity for the designer to explore design ideas. Incorporating structural design in the initial design phases can lead to more efficient and cost-effective structures, but in order to do so practical tools should be available to the designer to deepen the design space and allow them to efficiently explore it. Generative AI models could provide a solution to the problem and expand the design space by learning from the data provided. This thesis explores such a solution by means of training an AI model (Variational Autoencoder) on an artificial dataset of 2D lattice patterns. To do that, a dataset of 6 unique 2D lattice patterns is created and used to train a simple VAE model. The methodology used in this thesis is to validate if an AI can aid in the exploration of design ideas by providing a larger design space with newly generated data that contain learned features from the dataset rather than through constraints set by the designers. The results show that the VAE model can learn features and provide a greater diversity of design than the original dataset through newly generated designs. The output of the VAE model in this thesis is then explored for possible integrations into the design process for the early stages of design. For this, the generated patterns that are distinct and unique are identified and applied to a shell structure to explore topology design ideas.

Architects play a pivotal role play in the sustainable development of the built environment, shaping spaces that minimize environmental impact while enhancing human well-being. However, due to the escalating complexity of design requirements, designers are becoming underequipped to solve today’s design challenges. Solving such problems involves balancing multiple different co-dependent factors, such as spatial configuration, daylight satisfaction and embodied carbon. Currently this is addressed through inter-disciplinary collaboration where different parties focus on their own specialisation. As a result, a significant gap exists between design and building performance analysis leading to underperforming and expensive solutions. An artificial intelligence-powered platform could aid in the design process by bridging the gap between different disciplines. Such a platform could ensure that all aspects of a project are considered holistically, reducing the risk of conflicting decisions that cause lacking performance and unnecessary expenses. Simply consolidating information, as done with building information modelling (BIM), is not sufficient. Instead, the platform must provide performance-driven recommendations, offering deeper insights when conflicts between different disciplines arise. This work aims to develop a framework through which deep learning methods can be applied to create floorplans informed by performance-based criteria and user guidance. Since functionality is heavily dependent on site-conditions and requirements laid out by the client, the relationship between form and function must be uncovered dynamically. To achieve this, inspiration will be drawn from the field of reinforcement learning which allows the training of a neural network without the use of training data.

Architects play a pivotal role play in the sustainable development of the built environment, shaping spaces that minimize environmental impact while enhancing human well-being. However, due to the escalating complexity of design requirements, designers are becoming underequipped to solve today’s design challenges. Solving such problems involves balancing multiple different co-dependent factors, such as spatial configuration, daylight satisfaction and embodied carbon. Currently this is addressed through inter-disciplinary collaboration where different parties focus on their own specialisation. As a result, a significant gap exists between design and building performance analysis leading to underperforming and expensive solutions. An artificial intelligence-powered platform could aid in the design process by bridging the gap between different disciplines. Such a platform could ensure that all aspects of a project are considered holistically, reducing the risk of conflicting decisions that cause lacking performance and unnecessary expenses. Simply consolidating information, as done with building information modelling (BIM), is not sufficient. Instead, the platform must provide performance-driven recommendations, offering deeper insights when conflicts between different disciplines arise. This work aims to develop a framework through which deep learning methods can be applied to create floorplans informed by performance-based criteria and user guidance. Since functionality is heavily dependent on site-conditions and requirements laid out by the client, the relationship between form and function must be uncovered dynamically. To achieve this, inspiration will be drawn from the field of reinforcement learning which allows the training of a neural network without the use of training data.