Please use this identifier to cite or link to this item: https://www.um.edu.mt/library/oar/handle/123456789/79233
Title: Implementation and fine-tuning of the Big Bang-Big Crunch optimisation method for use in passive building design
Authors: Robic, Florian
Micallef, Daniel
Borg, Simon Paul
Ellul, Brian
Keywords: Sustainable buildings -- Design and construction
Artificial intelligence
Algorithms
Buildings -- Environmental engineering
Heating
Issue Date: 2020
Publisher: Elsevier BV
Citation: Robic, F., Micallef, D., Borg, S. P., & Ellul, B. (2020). Implementation and fine-tuning of the Big Bang-Big Crunch optimisation method for use in passive building design. Building and Environment, 173, 106731.
Abstract: Passive building design involves consideration of a number of design variables which can be optimised with respect to some design objective. The use of optimisation methods in building design is well established but the implementation of the Big Bang-Big Crunch algorithm to passive building design is still as yet unexplored in detail. The aim of this paper is to demonstrate the usefulness of the method as well as to fine-tune its characteristic parameters for its efficient use in building design applications. Two building design scenarios are used to have more than one test case and the number of design variables are limited to an extent that an enumeration search can be performed in a reasonable time frame from which the optimum can be precisely established. The Big Bang-Big Crunch algorithm is used and validated against the true solution from the enumeration search. Results show that the α and β parameters should be set to 0 and around 0.8 respectively. The ratio of the sample size to the number of design variables should ideally be between 2.6 and 2.9 for the algorithm to remain efficient and at the same time successful. The paper also discusses the computational efficiency gain of the Big Bang-Big Crunch algorithm compared to a full enumeration search. Computational savings of more than 90% are possible.
URI: https://www.um.edu.mt/library/oar/handle/123456789/79233
Appears in Collections:Scholarly Works - FacBenED



Items in OAR@UM are protected by copyright, with all rights reserved, unless otherwise indicated.