Keywords: commercial buildings, conceptual design, genetic algorithm, polygon partitioning.

The conceptual design of commercial buildings is a non-trivial problem because of the large number of variables and objectives results in huge and complex search spaces. To help the designers look for good solutions within the search space, several evolutionary decision support systems have been developed [1,2,3,4,5,6]. However these systems have so far been limited to buildings with rectangular floor plans.

This paper, presents a genetic algorithm based methodology capable of producing designs for buildings with an orthogonal boundary and voids that represents a significant improvement over existing research.

This work utilises polygon partitioning techniques to reduce a building floor plan to a collection of rectangular sections, which can then be solved independently. The partitioning of irregular polygons is a major topic within the field of computational geometry however most established partitioning techniques involve triangulation e.g. triangulation via monotone mountains. Unfortunately one of the most important features of commercial buildings is that columns must be arranged in rectangular grids. Therefore a sweep line algorithm has been developed that partitions the orthogonal boundary into rectangular sections which are then assigned column spacings. The genetic algorithm then exchanges column spacings between individuals, in a population of potential solutions, in order to search for suitable spacing arrangements. However to prevent unrealistic solutions being generated, a novel representation has also been developed to ensure that the evolutionary operations of the genetic algorithm are not too disruptive. This representation monitors the spacings of adjacent, rectangular sections and updates any adjacent sections when one is modified. This ensures that lines of columns do not fracture between sections in the building.

The number of initial variables has also been reduced, because this work is aimed at the conceptual design stage, so that the user need only specify the location and shape of the external boundary and any voids.

The work also uses XML to represent individual solutions to enable them to be exported to other analysis packages and also to allow for their visualisation as a virtual reality world using X3D. These X3D visualisations can be combined with server software designed by Blaxxun [7] to create a structure that can be viewed both independently and in collaboration over the Internet. This enables a distributed design team to view the structure in a collaborative virtual reality environment.

Overall the aim of this system is not to evolve an optimum structure but to search the solution space and return several distinct options to the user. These options can then be taken forward to the later design stages.

1

S. Khajehpour and D.E. Grierson, "Profitability versus safety of high-rise office buildings", Journal of Structural and Multidisciplinary Optimisation, 25, 1-15, 2003. doi:10.1007/s00158-003-0297-4

2

S. Khajehpour and D.E. Grierson, "Filtering of Pareto-Optimal trade-off surfaces for building conceptual design", in B.H.V. Topping & B. Kumar, (eds) Optimzation & Control in Civil & Structural Engineering, Civil-Comp Press, Edinburgh UK, 63-70, 1999. doi:10.4203/ccp.60.4.1

3

M.Y. Rafiq, Y. Bugmann and D.J. Easterbrook, "Building concept generation using genetic algorithms integrated with neural networks", Proceedings of AI in Structural Engineering, IT for Design, Manufacturing, Maintenance and Monitoring, EG-SEA-AI, Wierzba, Poland, 165-173, 1999.

4

I.C. Parmee, "Evolutionary and adaptive strategies for efficient search across whole system engineering design hierarchies", AIEDAM, 12, 431-445, 1998. doi:10.1017/S0890060498125039

5

J.C. Miles, G.M. Sisk and C.J. Moore, "The conceptual design of commercial buildings using a genetic algorithm", Computers and Structures, 79, 1583-1592, 2001. doi:10.1016/S0045-7949(01)00040-2

6

G.M. Sisk, J.C. Miles and C.J. Moore, "Designer centered development of GA-based DSS for conceptual design of buildings", Journal of Computing in Civil Engineering, 17(3), 159-166, 2003. doi:10.1061/(ASCE)0887-3801(2003)17:3(159)

7

Blaxxun Technologies, www.blaxxun.com, [Accessed 20 May 2005].

purchase the full-text of this paper (price £20)

go to the next paper
return to the table of contents
return to the book description
purchase this book (price £80 +P&P)