Shaper-GA: Automatic shape generation for modular house design

Authors

Abstract

This work presents a Genetic Algorithm (GA) approach to produce automatic designs for modular houses: Shaper-GA. A set of architectural design rules defining a language of design is incorporated into the GA fitness function. When possible genetic drift or local convergence might be occurring, the method starts an adaptive mutation rate to overcome fitness stagnation. The GA tool efficiently produces several layout solutions obeying the design rules and further placement constraints. Such a tool can be integrated into an appropriate user interface allowing future house owners or construction companies to customize their own house and answering the client's' requirements while complying with a language of design.

Keywords

Genetic Algorithms, Real-world problem, Module Placement Optimization

Subject

Genetic Algorithms

Conference

GECCO-The Genetic and Evolutionary Computation Conference 2018, July 2018

DOI