The Application of the IVCGA to Conceptual Building Design

Typically building design involves a number of individuals from a variety of disciplines, each with a differing emphasis on the design requirements. Thus, the objective of the study, presented in this section, is to illustrate how the IVCGA can be used in a multi-disciplinary design environment. The exercise has been conducted in close collaboration with an Architect, a Structural Engineer and a Heat and Ventilation Engineer, and it is noted that the differential requirements of each are:

• The Architect – Space, functionality and appearance and user comfort.

• The Structural Engineer – Building stability, strength and safety.

• The Heat and Ventilation Engineer – Mechanical, electrical, heating cooling and ventilation systems.

In this example, only two construction materials, concrete and steel, are included. Choice of suit­able construction material could sometimes be a determining factor in selecting design options.

rview IGAS. Total number of evaluations – 2100 of

File View Type Coordinates Axes Order IGAS

Original Variables

columnMaterial

Fig. 3. Visualising solutions in Variable and Objective Spaces.

In Figure 3, concepts generated for each construction material (steel and concrete) is distinctly separated. This is a clear advantage of visualisation tools, which allow designers to investigate each alternative construction material independently and assess the suitability of each alternative against the design/client requirements. In most figures presented in this section the information is presented in the ‘objective space’ rather than the ‘variable space’. This shows another distinct feature of the IVCGA is the ability to present data in either variable space or objective space. At times an objective space representation may be preferred to variable space when design variables are discrete variables.

To satisfy the architectural space requirements for this preliminary study, it is assumed that a major architectural requirement is the amount of available open space: A flexible floor area could be used by various clients with different functionality requirements throughout the life of the building. Therefore, there is a need to minimise the floor area lost due to columns and fire escapes. As the existence of columns in the floor makes the area around the columns restricted for use, it was decided to use aim2 unusable area for each column. In this study, a typical floor area of 1000 m2 is assumed.

Figure 4 shows a population of designs generated after the first 20 generations of the GA. The frames on the right hand side show two alternative floor layouts in two different construction ma­terials (steel and concrete). Both layouts satisfy the architectural requirement of minimum area loss. As the requirements suggests, fewer columns and other restrictive structural elements mean more letable floor area and hence more long term revenue.

The user is able to define which variables are plotted against each other, enabling designers to investigate any interrelationships between various design parameters, very quickly. The additional use of colour enables the user to emphasise the clusters that are important to the user whilst keeping other interesting data also available. For example, in Figure 4, concepts presented by cyan and magenta colours represent clusters of high fitness solutions.

Not only does the use of colour identify clusters of good designs, but also enables the designer to view individual solutions in the vicinity of these clusters in order to learn more about the inter­relationship between design parameters. A particular strength of the IVCGA, is the ability to zoom in on particular areas of the solution space and to conduct a more focussed search in that area.

Fig. 4. Alternative concepts in two different construction materials.

To illustrate this, Figure 5 shows a 2D representation of the designs generated by the GA, based on the structural requirements. The left hand frame presents the entire population generated by the GA. No restrictions or filtering of information, to discourage the generation of unfeasible solutions, have been implemented. The rectangle on the left hand frame is interactively drawn by the designer, using the mouse, around the vicinity of 1000 m2 required floor area, as was specified in the brief. The frame on the right hand side of Figure 5 shows only solutions which are enclosed inside this region. At this stage the system allows the designer to either view current designs generated by the GA within this region or conduct further runs the GA to populate this area with more designs. If this option is selected, the system automatically penalises any designs outside this region if generated by the GA operation. This feature of the system, which interactively and dynamically defines constraints to the problem, is a unique and very useful decision support tool.

A similar investigation could be conducted by each member of the design team to identify regions of suitable designs that best satisfy their individual disciplinary requirements. Finally, these areas are superimposed on each other to identify a mutually inclusive region that partially satisfies the require­ments of all disciplines involved in the design. The frame at the left hand of Figure 6 presents all designs generated during separate runs of the GA. The cluster coloured copper represents all design that satisfies energy requirements. The green cluster represents Structural Engineer requirements and the cyan colour cluster represents the preferred region by the Architect. The red circle is the

Ongmal Variable*

Overview ICAS. Tolal number of evaluation* – 2100 of

‘ Overview ICAS Total number of evaluation* – 2100 of

n# VW»lyp# – ОУСГ. КМ A.«r.*oer

Original Venable*

* .re

(a) All clusters only 3 clusters

Fig. 6. Mutually inclusive region for all three disciplines.

mutually inclusive region that partially satisfies the requirements of all parties involved. The frame at the right hand side of Figure 6 only shows clusters of suitable designs identified by the three disciplines independently.

Figure 7 present the above information in 2D and 3D scatter plots showing all design alternatives generated during separate runs of the GA. The cluster of designs coloured magenta represents the mutually inclusive region that partially satisfies the requirements of all parties involved.

Conclusions

The use of interactive visualisation undoubtedly gives confidence to the designer particularly at the conceptual stage of the design process by allowing her/him to conduct a human-led search in order to explore many possible solutions that satisfy design/client requirements. The IVCGA, is an effective decision support tool as it presents the designers with the opportunity to conduct search ‘inside’ and ‘outside’ the defined search and solution spaces and to evaluate the merit of each design generated,

Mutually inclusive region

ETWiffr

• in

Fig. 7. Mutually inclusive region for all three disciplines in 2d and 3D views.

in term of existing or new design requirements. Visualisation tools, such as the IVCGA, enhances the multi-disciplinary design team members’ understanding of the overall design issues and the consequences of the effect of changing one parameter or objectives on the overall design. Using visualisation tools it is possible to identify that region of the solution space that partially satisfies the requirements of all parties involved. Due to the fragmentation of activities between disciplines, agreeing on a compromise design framework presents difficulties in current multi disciplinary design practice. It is hoped that the use of tools such as IVCGA can facilitate better communication between parties involved in the design process.