No abstract

It is well accepted that search is an effective model for design. Newell and Simons' Human Information Processing model is foundational to this view. Designers use symbols and structures to express, store, off-load, recall, and manage their work. They mix general and detailed elements, organize their problem-space differently, seek ways to identify repetitive tasks, and utilize external media. An integral aspect of design-search is the comparison of alternatives, because the goal is usually to come close, if not fully satisfy, a set of requirements. Searching problem-spaces with currently available tools is challenging due to a number of issues related to creating and comparing alternative representations of one's thought process and outcome. In this paper, we present Alt.Text, a prototype that we developed to explore strategies for supporting design search. While Alt.Text only handles text-based documents, we believe that many of its features can be generalized to the domain of architectural design.

Many commercial environmental analysis tools support the evaluation of a building model based on parameters assigned in the design process. However interoperability issues between different data exchange formats hinder the iteration between design and analysis. Because the engineering calculations involved in analysis and evaluation are not integrated with architectural design parameters, evaluation takes place after the design is already defined, and analysis in real-time is not possible. In this project we integrate architectural design and engineering constraints to support design evolution and decision making by using a set of performance objectives. We propose a framework for coupling performance knowledge with generative synthesis to address multidisciplinary design challenges in the Architecture, Engineering, and Construction (AEC) industry. We develop a tool to support design evaluation based on performance criteria: energy consumption, comfort, and cost. Results show real-time information exchange as links between architectural geometry and engineering parameters. The outcomes of this research describe workflows and methods to evaluate alternative design proposals at early stages in real time.

Responses delivered by a generative synthesis system (GSS) vary between creative solutions and unusable outcomes. The type of GSS response is driven by many factors such as: the design context, designer's interpretation, implementation environments, design language, and the GSS composition, among many factors. In this paper, we describe a GSS framework to provide a recipe for delivering responses, which can be qualified as solutions. The framework focuses on GSS composition. It includes descriptions for: building blocks, components, and building strategy. The framework is informed by generative design literature and by our experimentation. We present the framework through: a brief background to GSS, metrics, building blocks, components, and building strategy. We also show an example of GSS implementation and offer a brief discussion.

No abstract