This paper proposes an interactive computational approach for designing the layout of timber column-slab structures in a material-saving and structural-performance aware manner. The generated column slab layout meets the serviceability limit state deflection criteria by integrating gradient-based optimization within an agent system approach. This was achieved by extending the ABxM framework with an interactive design framework termed Timber Column Slab (TCS) solver. This solver is underpinned by three data structures: Firstly, an agent system representing a column-slab system for layout exploration. The finite element (FE) model that computes the slab’s nodal displacements, which enables rapid serviceability checks at each agent-system update in a real-time loop. With the TCS extension, the agent-based model uses the numerical analysis results to refine the decision of the adaptive agents. The real-time interaction built within the agent-based model is responsible for the live user guidance and design software integration. The proposed methodology has been built with multiple scientific stacks and encapsulated into an open-source framework in a C# environment. The developed method is tested on a wide range of real building floor plans of different scales and typologies, demonstrating how effective it is compared to previous efforts.
%0 Book Section
%1 Udaykumar.2023
%A Udaykumar, Keerthana
%A Orozco, Luis
%A Krtschil, Anna
%A Menges, Achim
%A Knippers, Jan
%B Integration of Design and Fabrication
%C Melbourne, Australia
%D 2023
%E Xie, Yi Min
%E Burry, Jane
%E Lee, Ting Uei
%E Ma, Jiaming
%I International Association for Shell and Spatial Structures (IASS)
%K 2023 architecture column-slab gradient-based interactive itke knippers krtschil menges optimization orozco structure timber udaykumar
%P 1574--1584
%T Interactive Gradient-Based Optimization Method for Column-slab Structures
%X This paper proposes an interactive computational approach for designing the layout of timber column-slab structures in a material-saving and structural-performance aware manner. The generated column slab layout meets the serviceability limit state deflection criteria by integrating gradient-based optimization within an agent system approach. This was achieved by extending the ABxM framework with an interactive design framework termed Timber Column Slab (TCS) solver. This solver is underpinned by three data structures: Firstly, an agent system representing a column-slab system for layout exploration. The finite element (FE) model that computes the slab’s nodal displacements, which enables rapid serviceability checks at each agent-system update in a real-time loop. With the TCS extension, the agent-based model uses the numerical analysis results to refine the decision of the adaptive agents. The real-time interaction built within the agent-based model is responsible for the live user guidance and design software integration. The proposed methodology has been built with multiple scientific stacks and encapsulated into an open-source framework in a C# environment. The developed method is tested on a wide range of real building floor plans of different scales and typologies, demonstrating how effective it is compared to previous efforts.
@inbook{Udaykumar.2023,
abstract = {This paper proposes an interactive computational approach for designing the layout of timber column-slab structures in a material-saving and structural-performance aware manner. The generated column slab layout meets the serviceability limit state deflection criteria by integrating gradient-based optimization within an agent system approach. This was achieved by extending the ABxM framework with an interactive design framework termed Timber Column Slab (TCS) solver. This solver is underpinned by three data structures: Firstly, an agent system representing a column-slab system for layout exploration. The finite element (FE) model that computes the slab’s nodal displacements, which enables rapid serviceability checks at each agent-system update in a real-time loop. With the TCS extension, the agent-based model uses the numerical analysis results to refine the decision of the adaptive agents. The real-time interaction built within the agent-based model is responsible for the live user guidance and design software integration. The proposed methodology has been built with multiple scientific stacks and encapsulated into an open-source framework in a C# environment. The developed method is tested on a wide range of real building floor plans of different scales and typologies, demonstrating how effective it is compared to previous efforts.},
added-at = {2024-01-09T16:59:11.000+0100},
address = {Melbourne, Australia},
author = {Udaykumar, Keerthana and Orozco, Luis and Krtschil, Anna and Menges, Achim and Knippers, Jan},
biburl = {https://puma.ub.uni-stuttgart.de/bibtex/2aa3379904047d210a2a472be5648dd61/petraheim},
booktitle = {Integration of Design and Fabrication},
editor = {Xie, Yi Min and Burry, Jane and Lee, Ting Uei and Ma, Jiaming},
interhash = {7bd6028489edba7dff66cb9df5fb539d},
intrahash = {aa3379904047d210a2a472be5648dd61},
keywords = {2023 architecture column-slab gradient-based interactive itke knippers krtschil menges optimization orozco structure timber udaykumar},
language = {eng},
pages = {1574--1584},
publisher = {International Association for Shell and Spatial Structures (IASS)},
timestamp = {2024-01-09T16:59:11.000+0100},
title = {Interactive Gradient-Based Optimization Method for Column-slab Structures},
year = 2023
}
