Doulgerakis, A.; (2007) Genetic Programming + Unfolding Embryology in Automated Layout Planning. Masters thesis , UCL (University College London).

Preview

PDF 4981.pdf Download (4MB)

Abstract

Automated layout planning aims to the implementation of computational methods for the generation and the optimization of floor plans, considering the spatial configuration and the assignment of activities. Sophisticated strategies such as Genetic Algorithms have been implemented as heuristics of good solutions. However, the generative forces that derive from the social structures have been often neglected. This research aims to illustrate that the data that encode the layout’s social and cultural generative forces, can be implemented within an evolutionary system for the design of residential layouts. For that purpose a co-operative system was created, which is composed of a Genetic Programming algorithm and an agent-based unfolding embryology procedure that assigns activities to the spaces generated by the GP algorithm. The assignment of activities is a recursive process which follows instructions encoded as permeability graphs. Furthermore, the Ranking Sum Fitness evaluation method is proposed and applied for the achievement of multi-objective optimization. Its efficiency is tested against the Weighted-Sum Fitness function. The system’s results, both numerical and spatial, are compared to the results of a conventional evolutionary approach. This comparison showed that, in general, the proposed system can yield better solutions.

Download activity - last month

Export as JSONCSVXML

Download activity - last 12 months

Export as JSONCSVXML

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item