Configuration and parameterization of optimization frameworks for the computational support of design exploration can become an exclusive barrier for the adoption of such systems by engineers. This work addresses the problem of defining the elements that constitute a multiple-objective design optimization problem, that is, design variables, constants, objective functions, and constraint functions. In light of this, contributions are reviewed from the field of evolutionary design optimization with respect to their concrete implementation for design exploration. Machine learning and natural language processing are supposed to facilitate feasible approaches to the support of configuration and parameterization. Hence, the authors further review promising machine learning and natural language processing methods for automatic knowledge elicitation and formalization with respect to their implementation for evolutionary design optimization. These methods come from the fields of product attribute extraction, clustering of design solutions, relationship discovery, computation of objective functions, metamodeling, and design pattern extraction.

In this article, two methods to develop and optimize accompanying building spatial and structural designs are compared. The first, a coevolutionary method, applies deterministic procedures, inspired by realistic design processes, to cyclically add a suitable structural design to the input of a spatial design, evaluate and improve the structural design via the finite element method and topology optimization, adjust the spatial design according to the improved structural design, and modify the spatial design such that the initial spatial requirements are fulfilled. The second method uses a genetic algorithm that works on a population of accompanying building spatial and structural designs, using the finite element method for evaluation. If specific performance indicators and spatial requirements are used (i.e., total strain energy, spatial volume, and number of spaces), both methods provide optimized building designs; however, the coevolutionary method yields even better designs in a faster and more direct manner, whereas the genetic algorithm based method provides more design variants. Both methods show that collaborative design, for example, via design modification in one domain (here spatial) to optimize the design in another domain (here structural) can be as effective as monodisciplinary optimization; however, it may need adjustments to avoid the designs becoming progressively unrealistic. Designers are informed of the merits and disadvantages of design process simulation and design instance exploration, whereas scientists learn from a first fully operational and automated method for design process simulation, which is verified with a genetic algorithm and subject to future improvements and extensions in the community.

Nowadays, on the basis of significant work carried out, architectural adaption structures are considered to be intelligent entities, able to react to various internal or external influences. Their adaptive behavior can be examined in a digital or physical environment, generating a variety of alternative solutions or structural transformations. These are controlled through different computational approaches, ranging from interactive exploration ones, producing alternative emergent results, to automate optimization ones, resulting in acceptable fitting solutions. This paper examines the adaptive behavior of a kinetic structure, aiming to explore suitable solutions resulting in final appropriate shapes during the transformation process. A machine learning methodology that implements an artificial neural networks algorithm is integrated to the suggested structure. The latter is formed by units articulated together in a sequential composition consisting of primary soft mechanisms and secondary rigid components that are responsible for its reconfiguration and stiffness. A number of case studies that respond to unstructured environments are set as examples, to test the effectiveness of the proposed methodology to be used for handling a large number of input data and to optimize the complex and nonlinear transformation behavior of the kinetic system at the global level, as a result of the units’ local activation that influences nearby units in a chaotic and unpredictable manner.

This paper presents research in the development of heuristic evolutionary algorithms (EAs) for generating and exploring differentiated force-based structures. The algorithm is weighted toward design exploration of topological differentiation while including specific structural and material constraints. An embryological EA model is employed to “grow” networks of mass-spring elements achieving desired mesh densities that resolve themselves in tensile force (form-active) equilibrium. The primal quadrilateral quadrisection method serves as the foundation for a range of extensible subdivision methods. Unique to this research, the quad is addressed as a “cell” rather than a topological or geometric construct, allowing for the contents of the cell to vary in number of mass-spring elements and orientation. In this research, this approach has been termed the quadrilateral quadrisection with n variable topological transformation method. This research culminates with the introduction of a method for grafting meshes where emergent features from the evolved meshes can be transposed and replicated in an explicit yet informed manner. The EA and grafting methods function within a Java-based software called springFORM, developed in previous research, which utilizes a mass-spring based library for solving force equilibrium and allows for both active (manual) and algorithmic topology manipulation. In application to a specific complex tensile mesh, the design framework, which combines the generative EA and mesh grafting method, is shown to produce emergent and highly differentiated topological arrangements that negotiate the specific relationships among a desired maximal mesh density, geometric patterning, and equalized force distribution.

This paper presents the development of an optimization methodology for selecting the lowest monetary cost combinations of building technologies to meet set operational energy reduction targets. The new optimization algorithm introduced in this paper departs from the notion that optimal design choices over a large set of design parameters and properties can be driven by energy targets. We assume that design parameters are determined by many concurrent considerations fighting over the attention span of the design team. Our approach starts from a design outcome and asks the question, which set of discrete technologies are the right mix to reach an energy target in the cost optimal way? Such an approach has to face the challenge that the properties of market-available building technologies have a discrete nature that makes their optimal selection a combinatorial problem. The optimization algorithm searches the discrete combinatoric space by maximizing the following objective function: calculated energy savings divided by premium cost, where cost is defined as the additional cost over a baseline solution. The algorithm is codified into a custom MATLAB script and when compared to prescriptive methodologies is shown to be more cost effective and generically applicable given a palette of building technology alternatives and their corresponding cost data.

Performance assessment at early stages of buildings design is complicated by an inherent lack of information on the design and the uncertainty in how a building design may evolve to a final design. This pilot study reports on an initial quantification of such uncertainty associated with building energy performance and develops a method for informing decision makers of the risks in early design decisions under this uncertainty. Two case studies of building design decision situations under this uncertainty are explored along with using two different energy modeling tools: a reduced-order model and a high-order model. The intended contribution is to identify if a decision can be made with confidence in early design given a high level of uncertainty in the evolution of a design and what models can support decisions of this sort. Integration of the proposed decision support approach with a computer-aided design model is shown as well.

A platform for experimenting with population-based design exploration algorithms is presented, called Dexen. The platform has been developed in order to address the needs of two distinct groups of users loosely labeled as researchers and designers. Whereas the researchers group focuses on creating and testing customized toolkits, the designers group focuses on applying these toolkits in the design process. A platform is required that is scalable and extensible: scalable to allow computationally demanding population-based exploration algorithms to be executed on distributed hardware within reasonable time frames, and extensible to allow researchers to easily implement their own customized toolkits consisting of specialized algorithms and user interfaces. In order to address these requirements, a three-tier client–server system architecture has been used that separates data storage, domain logic, and presentation. This separation allows customized toolkits to be created for Dexen without requiring any changes to the data or logic tiers. In the logic tier, Dexen uses a programming model in which tasks only communicate through data objects stored in a key-value database. The paper ends with a case study experiment that uses a multicriteria evolutionary algorithm toolkit to explore alternative configurations for the massing and façade design of a large residential development. The parametric models for developing and evaluating design variants are described in detail. A population of design variants are evolved, a number of which are selected for further analysis. The case study demonstrates how evolutionary exploration methods can be applied to a complex design scenario without requiring any scripting.

Building performance simulation and genetic algorithms are powerful techniques for helping designers make better design decisions in architectural design optimization. However, they are very time consuming and require a significant amount of computing power. More time is needed when two techniques work together. This has become the primary impediment in applying design optimization to real-world projects. This study focuses on reducing the computing time in genetic algorithms when building simulation techniques are involved. In this study, we combine two techniques (offline simulation and divide and conquer) to effectively improve the run time in these architectural design optimization problems, utilizing architecture-specific domain knowledge. The improved methods are evaluated with a case study of a nursing unit design to minimize the nurses’ travel distance and maximize daylighting performance in patient rooms. Results show the computing time can be saved significantly during the simulation and optimization process.

Climate change, resource depletion, and worldwide urbanization feed the demand for more energy and resource-efficient buildings. Increasingly, architectural designers and consultants analyze building designs with easy-to-use simulation tools. To identify design alternatives with good performance, designers often turn to optimization methods. Randomized, metaheuristic methods such as genetic algorithms are popular in the architectural design field. However, are metaheuristics the best approach for architectural design problems that often are complex and ill defined? Metaheuristics may find solutions for well-defined problems, but they do not contribute to a better understanding of a complex design problem. This paper proposes surrogate-based optimization as a method that promotes understanding of the design problem. The surrogate method interpolates a mathematical model from data that relate design parameters to performance criteria. Designers can interact with this model to explore the approximate impact of changing design variables. We apply the radial basis function method, a specific type of surrogate model, to two architectural daylight optimization problems. These case studies, along with results from computational experiments, serve to discuss several advantages of surrogate models. First, surrogate models not only propose good solutions but also allow designers to address issues outside of the formulation of the optimization problem. Instead of accepting a solution presented by the optimization process, designers can improve their understanding of the design problem by interacting with the model. Second, a related advantage is that designers can quickly construct surrogate models from existing simulation results and other knowledge they might possess about the design problem. Designers can thus explore the impact of different evaluation criteria by constructing several models from the same set of data. They also can create models from approximate data and later refine them with more precise simulations. Third, surrogate-based methods typically find global optima orders of magnitude faster than genetic algorithms, especially when the evaluation of design variants requires time-intensive simulations.

Responding to growing concerns regarding energy-efficient facades, this paper describes the structure and process followed in the design of a responsive sun-shading system based on the use of rotating plates with two degrees of freedom. The proposal considers, among others, the definition of variable design parameters, areas of performance evaluation and control, and construction detailing development represented by a first 1:2 unit (module) model. In the process, computational simulation procedures were employed to explore configurational possibilities that would provide high-performance solutions to the light requirements of the particular covered spaces. In developing the system, it was noticed that due to the highly subjective requirements of users in terms of quantity and quality of lighting, a purely Boolean control system would not always be appropriate. Following from that, and taking advantage of the dynamic nature of the system, a further approach of control supported by fuzzy logic was also implemented at the operative state, whose logic is explained. Digital simulations were carried out to assess the performance of the system, and their results demonstrate more even light distribution levels compared to traditional systems.

The concept of shape grammars has often been proposed to improve or support creative design processes. Shape grammar implementations have the potential to both automate parts of the design process and allow exploration of design alternatives. In many of the existing implementations, the main focus is either on capturing the rationale of a particular existing grammar or on allowing designers to develop a new grammar. However, little attention is typically given to the actual representation of the design space that can be explored in the interface of the implementation. With such representation, a shape grammar implementation could properly support designers who are still in the process of designing and may not yet have a clear shape grammar in mind. In this article, an approach and a proof-of-concept software system is proposed for a shape grammar implementation that provides a visual and interactive way to support design space exploration in a creative design process. We describe the method by which this software system can be used and focus on how designers can interact with the exploration process. In particular, we point out how the proposed approach realizes several important amplification strategies to support design space exploration.