Generative Sustainable Performance Design Exploration Applied to Greek Urban Typologies

Abstract

Generative Design offers a paradigm for the automatic generation of variations of a design so that designers can quantify the bounds of possible Key Performance Indicator’s (KPI’s) and identify most promising variants based on their design objectives. While such studies are often performed to provide inspiration for high-end building designs, applying them on standard building typologies gives equally interesting insights on the impact of related design decisions. This paper presents a developed parametric model for the typical Greek Urban Typology, the Polykatoikia, and shows how this was utilized to identify sustainable design strategies for this typology. The methodology is applied on unfinished skeleton buildings in urban areas but can also be extended to new building designs. The KPI’s that the Generative Design algorithm optimizes towards are selected to fit important sustainability criteria, such as the minimization of summer solar loads and the maximization of winter solar load, together with standard User Experience Quality metrics related to Views and Daylight as well as standard leasable floor area maximisation criteria also taking into account standard dimensions of Structural core elements. The paper will also present opportunities for further optimization of the Polykatoikia design process, through the use of advanced design exploration and optimization algorithms and tools utilizing Machine Learning algorithms and advanced data visualisation.