Contextual optimisation workspace (COW) is a digital design tool under development that seeks to promote novel processes for the parametric design of architectural and engineering structures using multiple-objective optimisation (MOO) strategies based on evolutionary algorithms. Assembled within the Grasshopper (GH) graphical algorithm editor, the tool allows for constructive analytical comparisons between often conflicting aspects of a design that have historically been assumed incomparable, with the resulting compromises between objectives presented as iterations of a three-dimensional geometry with associated data. This increases the user’s capability to make informed decisions throughout the design process and to control the implementation of a potentially wide range of differing design objectives. This is useful in situations where an extended design team (more than one stakeholder) is collaborating on an architectural project, which would include most real-life applications. In such a team, all agents have different preferred outcomes. The effort to manage their expectations and fulfil their respective agendas becomes the aim of the resulting architectural scheme. This paper explores how the COW tool could benefit from the addition of components that simplify such decision-making processes, and showcases how such additions could be applied to the design of timber structures. Two new GH user objects were designed that simplify an extended and weighted control of MOO-based design actions using COW, while providing a mechanism that guarantees designs are not considered if they do not meet the minimal requirements set by the constraining frame conditions. Such stakeholder-based MOO designs are shown to be a beneficial addition to the COW system. It is argued that a more comprehensive version of this first attempt to allow differing desires to be used as a weighted part of the design process is a promising strategy for the design of future timber structures.
Keywords: COW, Grasshopper, timber structures, multiple-objective optimisation, evolutionary algorithms
Authors
Larsson M.
KTH Royal Institute of Technology. School of Architecture and the Built Environment. Dept. of Civil and Architectural Engineering. Div. of Building Materials. EnWoBio – Engineered Wood and Bio-based Building Materials Laboratory. Stockholm, Sweden
Wålinder M.
Login to download the PDF