Table Of ContentKATHOLIEKEUNIVERSITEITLEUVEN
FACULTEITTOEGEPASTEWETENSCHAPPEN
DEPARTEMENTARCHITECTUUR,STEDENBOUW
ENRUIMTELIJKEORDENING
KasteelvanArenberg,B-3001Leuven
In case of architectural design
Critique and praise of Case-Based Design in architecture
Jury: Proefschriftvoorgedragentot
VoorzitterProf.JoosVandewalle,vice-decaan hetbehalenvanhetdoctoraat
Prof.HermanNeuckermans,promotor indetoegepastewetenschappen
Prof.AndréLoeckx
Prof.JanSchreurs door
Prof.JohanWagemans
Prof.AlbertDupagne(UniversitédeLiège) AnnHeylighen
Prof.RivkaOxman(TechnionHaifa)
UDC72.011 Mei2000
KatholiekeUniversiteitLeuven–FaculteitToegepasteWetenschappen
KasteelvanArenberg,B-3001Leuven(Belgium)
Alle rechten voorbehouden. Niets uit deze uitgave mag worden vermenigvuldigd en/of
openbaar gemaakt door middel van druk, fotokopie, microfilm, elektronisch of op welke
anderewijzeookzondervoorafgaandetoestemmingvandeuitgever.
All rights reserved. No part of this publication may be reproduced in any form by print,
photoprint,microfilmoranyothermeanswithoutwrittenpermissionfromthepublisher.
D/2000/7515/17
ISBN90-5682-248-9
For mygrandmother
Acknowledgements
Five years ago, I had to decide on a subject for my graduate’s thesis. Professor
Herman Neuckermans made the suggestion to have a look at Case-Based Design, at
that time a ‘hot topic’ in the field of CAAD.Computers werenot exactlywhat Iwould
call my cup of tea, but I nevertheless accepted his suggestion – not in the last place
because it would enable me to spend one semester at the ETH in Zürich. Although the
thesis became an unqualified success in terms of scores, it left me with an uneasy
feeling. Judging from the literature, the applications of Case-Based Design in
architecture were not (yet) convincing. Yet, the underlying ideas seemed so plausible,
that I could not dismiss it as the umpteenth flash in the CAAD pan. Immediatelyafter
my graduation, I received a grant from the Fund for Scientific Research (FWO)
Flanders,whichenabledmetopursuethematterwithadoctoral research.
During the course of this research, I’ve had a lot of help and support. It would be
impossible to mention everyone I met and exchanged opinions with, who gave me
suggestions and encouragement. But among them, I feel a special obligation to thank,
above all, Professor Neuckermans, who generously invited me into his CAAD-lab,
thus givingme access tohis personal expertise,his researchbudget (especiallythepart
meant for travelling), his libraryand – last but not least – the dailyhelp and friendship
of his other CAAD-assistants. Among them, a very special thanks goes to Benjamin
Geebelen, Bart Geeraerts and Kris Nuyts for the 1013 ways in which they have
contributed to – and once in a while kept me from – my work. In 1997/1998, the
CAAD-lab was temporarily reinforced by Raf Segers, the first thesis student I’ve
supervisedand,moreimportantly,thelivingdynamobehindDYNAMO.
The data analysis for the evaluation of DYNAMO and the experimental study on
the effects of examples was performed in close collaboration with Dr. Ilse Verstijnen
from the department of Psychology. Collaborating with the “dépendance in de
Tiensestraat” – Professor Johan Wagemans, Ilse and Pim – was not just instructive, I
reallyenjoyedit.
The collaboration with Jan Bouwen from the department of Applied Economics
was less structural, and above all less self-evident, but therefore not less fruitful. I am
indebted to him for his exchanged opinions, suggestions and encouragement in the
course of my research, and for his detailed constructive comments on earlier drafts of
this text.
In addition to Professor Wagemans, I would like to thank Professor André Loeckx
and Professor Jan Schreurs for the stern and critical, but productive wayin which they
accomplished their task as members of the supervising committee. Thanks also to
Professor Albert Dupagne (Université de Liège) and Professor Rivka Oxman
(Technion,Haifa),whohaveacceptedtobeonthejury.
I am, of course, extremely grateful to the architects I’ve interviewed – Jan Delrue,
MauroPoponcini,Paul VanAerschot,WernerVandermeersch,Paul Vermeulenand
Peggy Winkels. They have given their precious time very generously and have made
thewritingofchapter4afascinatingandthoroughlyrewardingexperience.
In addition to the CAAD-colleagues, I was given considerable help (and support)
by many other staff members and students of our department. In particular, I would
like to thank all those who have contributed to the experiments. They are (amongst
others): the students and the titular of the 2nd year Basic Design course; the studio
teachers Evi Corne, (once again) Mauro Poponcini, Ann Verdonck and Hans
Verplancke; the external judges Guido Geenen and Ivo Vanhamme; the students and
the titular of the 4th year design studio; the studio teachers Hans De Petter, Leo Van
Broeck and (once again) Ivo Vanhamme; and the external judges Heike Löhmann and
Piet Stevens.
Karen Depoortere and Paul McHale were continuously helpful during my work by
keepinganeyeonthegrammarandspellingofthetext.
For encouraging me all the while, not only as I worked on this thesis, but also
during my study, I am very grateful to my parents and my sister and, for his daily on-
line support, to Toon. They saw me through the good times and the bad, and were
always convincedthat Iwouldsucceed,especiallywhenIwas not.
Thanks are due to many friends who have supported me in multiple ways.
However, particular thanks go to Liesbet, Veerle and Gunther for their extra
encouragement during the last months, and to the members of the working group
‘WomanandUniversity’,forholdingupthemselves as stimulatingexample.
I dedicate this work to my grandmother, who had to miss me on many Sunday
afternoons. My talents are hers, the opportunities I was given are the ones she should
havegot.
In caseof architectural design
CritiqueandpraiseofCase-BasedDesigninarchitecture
Architects are said to learn design byexperience. Learning design byexperience is the
essence of Case-Based Design (CBD), a sub-domain of Artificial Intelligence. Part I
critically explores the CBD approach from an architectural point of view, tracing its
origins in the Theory of Dynamic Memory and highlighting its potential for
architectural design. Seven CBD systems are analysed, experienced architects and
design teachers are interviewed, and an experiment is carried out to examine how
cases affect the design performance of architecture students. The results of this
exploration show that despite its sound view on how architects acquire (design)
knowledge, CBD is limited in important respects: it reduces architectural design to
problem solving, is difficult to implement and has to contend with prejudices among
the target group. With a view to stretching these limits, part II covers the design,
implementation and evaluation of DYNAMO (Dynamic Architectural Memory On-
line). This Web-based design tool tailors the CBD approach to the complexity of
architectural design by effecting three transformations: extending the concern with
design products towards design processes, turning static case bases into dynamic
memories and upgrading users from passive case consumers to active case-based
designers.
Case-Based Design in architectuur
Eenkritischestudie
Architecten, zo beweert men, leren ontwerpen uit ervaring. Leren ontwerpen uit
ervaring is de essentie van Case-Based Design (CBD), een sub-domein van de
Artificiële Intelligentie. In deel I wordt de CBD-benadering onderworpen aan een
architectuurkritische studie. CBD is gebaseerd op de Theorie van het Dynamische
Geheugen, en biedt het architectuurontwerpen heel wat voordelen. Zeven CBD
systemen worden doorgelicht, ervaren architecten en ontwerpbegeleiders worden
geïnterviewd, en de impact van concrete ontwerpervaring (cases) op de
ontwerpprestatie van architectuurstudenten wordt experimenteel onderzocht. Hieruit
blijkt dat CBD bijdraagt tot een beter inzicht in hoe architecten leren ontwerpen, maar
beperkt is qua visie op het architectuurontwerpen, implementatie en
gebruikerscontext. Deel II beschrijft het ontwerp, de implementatie en de evaluatie
van een Dynamisch Architectuur-Geheugen On-line (DYNAMO), een
ontwerpinstrument dat de CBD-aanpak poogt af te stemmen op de complexiteit van
het architectuurontwerpen. Het verschuift de aandacht van ontwerpproducten naar
ontwerpprocessen, vervangt statische case bases door een dynamisch geheugen, en
maakt van de gebruikers actieve case-producenten in plaats van passieve case-
consumenten.
CONTENTS
Introduction 1
Chapter1:Conceivingarchitecture
1.1Thecomplexityofarchitectural design 9
1.2Knowledgeengineeringinarchitectural design 11
1.2.1Thingstoknow 11
1.2.2Waysofknowing 12
1.3Knowledgechannels forconceivingarchitecture 14
1.3.1Heuristics 16
1.3.2Analogy 17
1.3.2.1Iconicanalogy 19
1.3.2.2Canonicanalogy 21
1.3.3Fromanalogytometaphor(andbacktoanalogy) 22
1.3.4Typeinanutshell 25
1.3.5Typesandinstances 30
1.3.6Agridacrossacomplexfield 34
1.4Priornotices 36
PartI.Acritical exploration of Case-Based Design in architecture
Chapter2:Whatis Case-Based Design based on?
FromArtificial Intelligenceover DynamicMemorytoCase-BasedDesign
2.1Artificial Intelligence 42
2.1.1What? 42
2.1.2Why? 43
2.1.3How? 43
2.2Case-BasedReasoning: theory,model andapplication 44
2.2.1TheTheoryofDynamicMemory 45
2.2.1.1Scripts 45
2.2.1.2Reminding 46
2.2.1.3Remindingandproblemsolving 46
2.2.2ThecognitivemodelbehindCBR 47
2.2.2.1Structureandorganisationofknowledge 47
2.2.2.2Reasoningprocesses 48
2.2.2.3Learning 50
2.2.2.4Generalisedknowledge 50
2.2.3Applications 51
2.2.3.1ConstructingCase-BasedReasoners 51
2.2.3.2Investigatinghumanintelligence 52
2.3.Case-BasedDesign: premiseandpromises 53
2.3.1Premise 54
2.3.2PromisesofCBDforarchitecture 55
2.4Summary 58
Chapter3:Acasebaseof Case-Based Design tools forarchitecture
3.1Archie-II 61
3.1.1Casebase 61
3.1.1.1Designcase 61
3.1.1.2Caserepresentation 62
3.1.1.3Memoryorganisation 63
3.1.2Retrieval 64
3.1.3Manipulation 64
3.1.4Discussion 64
3.1.5References 65
3.2CADRE 66
3.2.1Casebase 66
3.2.1.1Designcase 67
3.2.1.2Caserepresentation 67
3.2.1.3Memoryorganisation 68
3.2.2Retrieval 68
3.2.3Manipulation 68
3.2.3.1Insertion 68
3.2.3.2Dimensionaladaptation 68
3.2.3.3Topologicaladaptation 69
3.2.3.4Evaluationandvisualinspection 70
3.2.4Discussion 70
3.2.5References 70
3.3FABEL 72
3.3.1Casebase(s) 72
3.3.1.1Designcase 72
3.3.1.2Caserepresentation 72
3.3.1.3Memoryorganisation 73
3.3.2Retrieval 73
3.3.3Manipulation 75
3.3.4FurtherFABELingredients 75
3.3.5Discussion 76
3.3.6References 77
3.4IDIOM 78
3.4.1Casebase 78
3.4.1.1Designcase 78
3.4.1.2Caserepresentation 78
3.4.1.3Memoryorganisation 79
3.4.2Retrieval 79
3.4.3Manipulation 79
3.4.4Discussion 80
3.4.5References 81
3.5PRECEDENTS 83
3.5.1Casebase 83
3.5.1.1Designcase 84
3.5.1.2Caserepresentation 84
3.5.1.3Memoryorganisation 85
3.5.2Retrieval 86
3.5.3Manipulation 87
3.5.4Discussion 87
3.5.5References 88
Description:CAAD-lab was temporarily reinforced by Raf Segers, the first thesis student I've The data analysis for the evaluation of DYNAMO and the experimental was less structural, and above all less self-evident, but therefore not less geïnterviewd, en de impact van concrete ontwerpervaring (cases) op de