European Centre for Living Technology

Evolutionary Building Design

by

Prof. Jonathan Wright, Loughborough University, UK

Abstract

Buildings are complex and dynamic artefacts, the design of which is driven by the need to minimize costs whilst meeting client and occupant expectations (buildings are important to us – we spend 90% of our lives living and working inside them). With 50% of carbon emissions resulting from energy demand in buildings, and the need to limit anthropogenic climate change, the challenge of designing cost effective high performance buildings has never been greater.

The majority of building optimization problems can be classified as being mixed-integer problems having non-linear objective and constraint functions; some aspects of building design can also be classified as being topological (combinatorial) optimization problems. In this seminar, Jonathan Wright will describe the application of Evolutionary Algorithms and thermal performance modelling to the solution of four example building optimization problems:

1.    - the optimization of window geometry (binary variables, non-linear objective and constraints);

2.    - the simultaneous optimization of building fabric construction, air-conditioning system capacity, and the system control strategy (formulated as a mixed-integer, multi-objective, non-linearly constrained optimization problem);

3.    - the topological optimization of an air-conditioning system configuration (formulated as a mixed-integer, single-objective, highly constrained combinatorial problem);

4.    - the optimization of building layout (formulated as a mixed-integer single objective problem with constraints used to enforce desired topological connections between zones).

Jonathan Wright is Professor of Building Optimization at Loughborough University in the UK.