Learning while designing

Abstract

This paper describes how a computational system for designing can learn useful, reusable, generalized search strategy rules from its own experience of designing. It can then apply this experience to transform the design process from search based~knowledge lean! to knowledge based~knowledge rich!. The domain of application is the design of spatial layouts for architectural design. The processes of designing and learning are tightly coupled.

A Framework for Situated Design Optimization

“…This leads to better problem formulation in terms of reduced search spaces and improved starting points constraint functions of a future optimization task may be neglected in the tool. Nath and Gero (2004) use machine learning to let a system acquire strategic knowledge as mappings between past design contexts and design decisions that led to useful results. These mappings are then available for the system to achieve solutions to similar design tasks more efficiently.…”

Learning Symbolic Formulations in Design Optimization

“…Ellman et al developed a tool allowing the user to formulate, test, reformulate and visualize a tree of optimization strategies constructed by the user that may be applied onto test problems in order to identify relevant optimization strategies for particular design domains [11]. Nath and Gero developed a machine learning tool that acquires strategies as mappings between past design contexts and design decisions that led to useful results [2]. Campbell et al developed an automated synthesis tool called A-Design.…”

“…the learning and doing of the problem develop together [2]; the solutions produced by the symbolic design model are only as valid as the modeling assumptions that are built into the model, making "reasonable and good" modeling assumptions essential for a valid solution [3]; the adopted modeling formalism is fundamentally related to the kind of algorithms that the designer intends to use to solve the problem. Problem formulation and solution method selection co-evolve [1].…”

Collective Intelligence: An Agent-Based Approach to Programmatic Organization in Architecture

“…Mitchell [2], for example, discusses the merits of automated design in architecture and also offers a method for generating least cost floor plan layouts [3]. In the past few decades various computational tools have been developed for automated design of floor plan layouts [4][5][8][9][10][11][12][13][14][15][16][17][18]. In general, these tools attempted to find the best possible floor plan layout to a given objective and set of constraints by searching through the space of all possible solutions.…”