Comparing two evolutionary algorithm based methods for layout generation: Dense packing versus subdivision. Issue 3 (22nd July 2014)

• Record Type: Journal Article
• Authors: Koenig, Reinhard; Knecht, Katja
• Editors: Campbell, Matthew I.; Shea, Kristina
• Is Part Of: AI EDAM. Volume 28:Issue 3(2014)
• Page Start: 285
• Page End: 299
• Publication Date: 2014-07-22
• Subjects: Dense Packing, -- Evolutionary Algorithm, -- Layout Generation, -- Subdivision; Engineering design -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; 620.00420285
• DOI: 10.1017/S0890060414000237
• Languages: English
• ISSNs: 0890-0604

Abstract

Abstract: We present and compare two evolutionary algorithm based methods for rectangular architectural layout generation: dense packing and subdivision algorithms. We analyze the characteristics of the two methods on the basis of three floor plan scenarios. Our analyses include the speed with which solutions are generated, the reliability with which optimal solutions can be found, and the number of different solutions that can be found overall. In a following step, we discuss the methods with respect to their different user interaction capabilities. In addition, we show that each method has the capability to generate more complex L-shaped layouts. Finally, we conclude that neither of the methods is superior but that each of them is suitable for use in distinct application scenarios because of its different properties.