id 
ddss9487 
authors 
Snijder, H.P.S. 
year 
1994 
title 
The Use of Genetic Algorithms in Spatial Optimisation Problems 
source 
Second Design and Decision Support Systems in Architecture & Urban Planning (Vaals, the Netherlands), August 1519, 1994 
summary 
The manipulation of a set of associative data usually involves the search of a huge searchspace (e.g. a set of 20 elements can be ordered in 20! ways, which is approximately equal to 2.4e+ 18). Rooms on a floor can be considered as a set of associative data. Optimising such a set according to some criterion (for example, minimising the distance between the related elements) can therefore be a daunting task. In order to assist in this task, a program (called ROP) has been developed, which graphically represents the relations in a matrix. The points in this matrix can be moved manually, thereby transforming the search process into a visual task. However, a considerable amount of skill remains required. In order to further alleviate the user in this task, ROP has been augmented with a Genetic Algorithm. A genetic algorithm is ideally suited to deal with very large searchspaces, and proved to be a valuable addition to ROP. In addition to employing the genetic algorithm for finding the optimal ordering, it can also be made to suggest several different orderings with approximately equal fitness, thereby providing elementary creativity support. The combination of ROP with a genetic algorithm provides a generic tool for the manipulation of all multivariate or associative data sets; in as well as outside the design realm. 
series 
DDSS 
references 
Contenttype: text/plain

last changed 
2003/08/07 14:36 
