CumInCAD is a Cumulative Index about publications in Computer Aided Architectural Design
supported by the sibling associations ACADIA, CAADRIA, eCAADe, SIGraDi, ASCAAD and CAAD futures

PDF papers
id a126
authors Finucane E, Derix C and Coates P
year 2006
title Evolving Urban Structures using Computational Optimisation
source Proceedings of the Generative Arts conference, Milan, 2006
summary This paper investigates the use of computer analogies for naturally inspired optimisation techniques as an aid to developing the site layout and massing for the new World Trade Centre development in Pristina Kosovo, which is being designed and developed by 4M Group architectural company, in conjunction with the Advanced Modelling Group Aedas. The development of a genetic algorithm will incorporate various techniques, that have been developed in the field of multi-objective optimisation, to create three dimensional massing models, and site layout solutions which partially fulfil the Prisina brief requirements, which are taken from specifications created by 4M Group. Genetic algorithms are based on natural evolutionary principles which are explained in this paper. It will incorporate Pareto concepts to manage the optimisation of the various objective functions. For example, these will include volume and position of units, which will ensure that the different and sometime conflicting needs of the site are balanced throughout the optimisation. This type of problem is often known as an NP-complete (non-determinate polynomial time) problem. This will provide architects and planners with a number of Pareto optimised site massing solutions as an aid to the design process. An initial investigation into the specifics of the Pristina site requirements, will be followed by an investigation into the the genetic algorithm which is created in Visual Basic for Applications (VBA) linked with AutoCAD as the graphical output of the code. The embryology (development) of the various solutions from the genetic information incorporates an Ďantí pheromone trail model, which simulates the action of ants during food foraging, as a tool for initial route planning within the site. Diffusion and cellular automata are used during the development of the solution to construct the massing for the site.
keywords urban planning, evolutionary algorithms, pareto optimization, Lindenmayer systems, ant-colony optimization, cellular automaton
series other
type normal paper
full text file.pdf (3,004,130 bytes)
references Content-type: text/plain
last changed 2012/09/20 16:33