Computational Design Sustainability: A Conceptual Framework for Built Environment Research Sahar Soltani1, Gabriela Dias Guimaraes2, Pan Liao3, Victor Calixto4, Ning Gu5 1,2,3,4,5University of South Australia, UniSA Creative, IVE: Australian Research Centre for Interactive and Virtual Environments, SA, 5000, Australia 1,2,3,4{sahar.soltani|gabriela.dias_guimaraes|pan.liao|victor.calixto}@mymail. unisa.edu.au [email protected] This paper presents and demonstrates a ``Computational Design Sustainability'' (CDS) framework, inspired by ``Computational Sustainability'' (CS), which is a new area in computational research (C. Gomes & Yang, 2011). CS aims to apply techniques from computer science to address sustainability problems affecting a wide range of fields from environmental sciences to social studies. While CS has been broadly embraced in environmental science, the great potential of this concept to address grand challenges and solve complex problems seems to have not been adequately explored in the built environment domain. Therefore, this paper attempts to formally investigate the application of CS in built environment research addressing different scales of design problems with computational design through proposing the new concept of CDS . These approaches are demonstrated and evaluated through a range of projects collectively conducted by the research team. CDS proposes to advance computational deign research by creating a trade-off between pillars of sustainability in an integrated multifaceted and multidisciplinary approach. The presented conceptual framework provides a formal means to critically understand and further advance these approaches in a systematic way suitable for future development and broader application. Keywords: Computational Design Sustainability, Computational Sustainability, Computational Design, Sustainable Development, Built Environment Research INTRODUCTION long-term goal shifted from growth implies an in- As a result of high consumption of natural resources crease of quantity or size, to development, that sug- and increased CO2 emissions in the last century, the gests an improvement in quality (Costanza, 2020). human subsystem can be considered a major com- This rapid change of contemporary, social and ponent of the overall ecosystem, so much that a new urban scheme, concurrent with the responsibility of geologic epoch called “Anthropocene” has started issues related to sustainability intensifies the need to (Daly, 2005; Steffen et al., 2007). In this context, the develop solutions and adapt to the ongoing changes D1.T1.S2. DESIGN AND COMPUTATION OF URBAN AND LOCAL SYSTEMS – XS to XL - Volume 1 - eCAADe 38 | 219 (Fox, 2009). In the built environment domain, this methods in computational design can benefit sus- calls for innovative approaches and technologies that tainability. Through a critical literature review, the enable a shift from individual to collaborative, disci- paper will focus on the key terms and issues of the plinary to interdisciplinary, and implicitly to explic- topic considering different scales of built environ- itly managed design processes (Haidar, Underwood, ment research, varying from individual buildings to & Coates, 2019). Such design processes attempt to urban scales. The presented conceptual framework come up with scalable solutions beyond mere com- provides a formal means to critically understand and putational processes, inspired by real-world prob- further advance these approaches in a systematic lems (C. P. Gomes, 2009). way suitable for future development and broader Computational sustainability (CS) adopts tech- application. In other words, it can assist in identi- niques from computer sciences to address sustain- fying new research directions, methodological ap- ability problems (C. Gomes & Yang, 2011) while af- proaches, potential applications (i.e. analysis, collab- fecting a wide range of areas, from environmental oration, management or generation), and also find- science to social studies, only recently started to ex- ing the available computational techniques connect- plore fields such as smart cities development (Simon- ing to CD. This framework is demonstrated and eval- detti & Birch, 2017). In the design and built envi- uated through a range of projects collectively con- ronment domain, emerging computational method- ducted by the research team. ologies, such as parametric design, are indeed rep- resenting the potential for supporting sustainability. COMPUTATIONAL SUSTAINABILITY However, considering the speedy advancement of Computational Sustainability (CS) (C. Gomes & Yang, technology and its impact on computational sustain- 2011) is a novel research field that attempts to em- ability, such potential to address grand challenges ploy computer science technologies such as Artificial and solve complex problems seems to have not been Intelligence and Machine Learning to address broad adequately explored in an integrated, collaborative sustainability problems integrating environmental, and multidisciplinary approach. This approach can economic and social dimensions. These problems in be explored in Computational Design (CD) due to in- CS are classified according to the relevant computer herent nature to engaging computational thinking as application that is utilized to address the problems (C. well as innovative technological capacities (Celani & Gomes & Yang, 2011). As such, common categories of Veloso, 2017; Oxman, 2006). CD has been altering the the research field in the CS framework includes com- traditional ways of conducting research and design putational simulation, constraint optimization and activities while addressing multi-faceted complexed machine learning, data mining, planning control and issues in the built environment. scheduling, policy and action learning. Since the na- Therefore, in this study, we apply and extend ture of such problems is complex and concerns var- CS in a design context to explore the question “how ious disciplines, they often require multidisciplinary computational design approaches in built environ- approaches (Silva & Analide, 2019). Computation ment research enable and advance sustainability?” and sustainability arguably move each other forward Mapping the two concepts of CS and CD, the aim of while adapting methods and mathematical models this paper is to introduce, define, and demonstrate developed to solve one problem to tackle another “Computational Design Sustainability” focusing on problem, even transferable and scalable across var- the key conceptual components, methodological ap- ious disciplines (Eaton, Gomes, & Williams, 2014; C. proaches, and implementations within the built en- Gomes, 2019). This framework has been mapped vironment discipline. More specifically, we are in- to different problems solvable by various computa- terested in formally understanding how the applied tional methods. 220 | eCAADe 38 - D1.T1.S2. DESIGN AND COMPUTATION OF URBAN AND LOCAL SYSTEMS – XS to XL - Volume 1 As an example, many computational approaches analysis, interaction, aesthetic, and presentation. The have been developed to tackle issues that involve term is closely related to the earliest understand- environmental concerns and resource management ing of Computer-Aided-Architectural-Design (CAAD) with economic and corresponding social problems (Mitchell, 1975) and Design Computing (Gero & Ma- such as poverty. Rapid population growth concur- her, 1999), refereeing to process computing infor- rent with the high-speed consumption of natural re- mation towards a design. It can be wrongly asso- sources has amplified the need for solutions to tackle ciated with Computer-Aided Design (CAD) systems the growing problem of poverty. However, lack as a computerisation process (Terzidis, 2006), or in of data and high cost of the available social data other words, as digital drafting tables. Computeri- such as census data has triggered motives for CS re- sation and Computation differ in the design process, searchers to employ machine learning methods and where the first is the literal translation of the con- using satellite and remote sensing data technologies ceived design to a new platform, which means not to extract socioeconomic data in a less costly and altering the amount of information produced during faster way. This approach has facilitated the process the process. The latter establishes interactions be- of data collection in some countries where access tween information and design elements through an to such data sources is difficult, demonstrating the algorithmic process, by increasing the amount of in- effectiveness of computational and multidisciplinary formation that was given as an input, contributing to methods in addressing sustainable issues (C. Gomes, adding new information to the design process. 2019). In the last decades, with the advent of embodied While CS has been embraced in environmental script languages in Computer-aided Design (CAD) science and is being broadly applied across other and Building Information Modelling (BIM) software disciplines, the effective application of the avail- as well as their visual editor of algorithms, such as able computational techniques and tools to integrate Grasshopper and Dynamo, CD became popularised sustainability in built environment design research, among architects and scholars (Burry, 2013; Celani however, remained unclear. Such methodologies & Vaz, 2012; Oxman, 2006). This has empowered can be explored in Computational Design (CD) due designers to do scripts in a more user-friendly