OPTIMIZATION OF OFFICE BUILDING FORM USING EVOLUTIONARY ALGORITHMS

Abstract

This dissertation explores some key strategies on form optimization of an architectural space using evolutionary algorithms to improve the performance of the space within. The architectural space taken into consideration is an office building since office buildings are one of the prime energy consumers of the built environment (BCA, 2010), (Cholakis, 2008). This particular exploration is demonstrated with the help of an experiment.

Form is usually determined by various parameters which include plot ratio, heights, orientation and planning profile. Manually handling all these parameters and evaluation criteria for various design iterations is difficult. Thus an evolutionary algorithm is used to link all these parameters and evaluation criteria in organized procedural stages of development, evaluation and feedback. Development involves the creation of a parametric model suitable for the experiment. Evaluation involves calculating the performance score for the parametric model with respect to the evaluation criteria. Feedback involves ranking of the designs based on their performance scores towards the end of the experiment.

An Evolutionary algorithm is a set of instructions following the evolutionary path of nature which can be applied to any optimization problem. There are various advantages of evolutionary algorithms which include its potential to simultaneously handle multiple diverse factors influencing the problem. It also returns with the performance feedback for all the iterations. These advantages of evolutionary algorithms make it a potential tool for architectural optimisation. Daylight and solar insolation have been taken as key evaluation criteria in the process of optimization since both these criteria play an important role in determining the performance of the form and the space within. The evaluation criteria are conflicting in nature, interdependent on each other in opposite directions. (Edwards and Tordellini, 2002, p.9-17), (Inkarojrit, n.d), (Seaton, 1989) Other evaluation criteria such as structural optimization are beyond the scope of the current exploration. However, it is noted that such evaluation criteria could be included.