Intelligent design objects (IDO): A cognitive approach for performance‐based design
Başlık çevirisi mevcut değil.
- Tez No: 523402
- Danışmanlar: Prof. Dr. ÖZER ÇİFTÇİOĞLU
- Tez Türü: Doktora
- Konular: Bilim ve Teknoloji, Mimarlık, Science and Technology, Architecture
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2009
- Dil: İngilizce
- Üniversite: Technische Universiteit Delft (Delft University of Technology)
- Enstitü: Yurtdışı Enstitü
- Ana Bilim Dalı: Belirtilmemiş.
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: 248
Özet
Özet yok.
Özet (Çeviri)
Architectural designs have far reaching impact on our lives. They satisfy our basic needs for safety, shelter and social interaction. Beyond that they address our need to live in an aesthetically appealing environment, and they form a contribution to our culture. This implies that a building should satisfy multiple criteria at the same time. It should be functional, aesthetic, energy efficient, cost effective, etc. As some criteria are conflicting, a design does not fulfil each criterion totally at the same time. Therefore we tolerate to some extend that they are not totally fulfilled. Also some requirements refer to non‐technical aspects of design, such as visual perception aspects. Such requirements are difficult to grasp, i.e. they are soft. These issues make it problematic to assess the overall suitability of a design. It is difficult to be conscious of the softness of design aspects and their manifold interdependencies to make such an assessment with certainty. Next to that the solution options are manifold. A vast amount of potential solutions should be investigated to be sure we did not miss a most favourable one. This is especially important because a slight variation of a certain design parameter may influence many criteria at the same time, and drastically improve or worsen the design. However, our time is limited to experiment with the design parameters. These factors make architectural design a complex task. The complexity may lead to designs that do not satisfy the criteria to a desirable extent. Unexpected problems occur in late phases of a design when they are costly to correct. Even they may be discovered after the erection of a building. This becomes more and more severe as the pressure to use resources with care increases. The motivation of the present thesis is to help architects to deal with the complexity of their task. In particular the goal is to provide means to reach designs that match multiple, soft criteria with increased certainty. This is accomplished by using advanced computational means that take care of several aspects of the design complexity. In particular the design objects, i.e. the physical elements that constitute a design are equipped with a reasoning mechanism. This way they 'know themselves' what to do to satisfy the architect's requirements. This means they exhibit intelligent behaviour. Using intelligent design objects an architect is able to focus on higher‐level aspects of a design. Such aspects are whether the priorities among criteria are adequate, and whether the right criteria are applied. The objects take care of the primary criteria put forward by an architect, so that one can exercise preferences with great awareness, being certain that solutions match the criteria. This approach is novel in particular as it is able to deal with soft issues in design. Existing computational methods, including the methods from the domain of classical artificial intelligence, are unable to deal with these issues. For example the novel approach enables a computer to handle demands for 'high visual privacy' or 'high functionality.' Such issues are conventionally considered to be incomputable, so that they form a bottleneck in evaluating the performance of a design. The advanced computational means used in the thesis to deal with such matters are known as soft computing methodologies. They are especially equipped to handle the complexity of design that is due to the softness of design criteria. In particular visual perception related aspects are vaguely understood and difficult to treat. This is handled in the approach by using a probabilistic visual perception model. The model is employed to assess perceptual properties of spaces with precision. Based on the results subtle perceptual differences can be weighed against other criteria, ensuring that they are taken into account together commensurately. The latter process is accomplished by using a special model for performance evaluation. It takes the softness of criteria into account during evaluation. Based on the results from the evaluation an algorithm is employed that generates design alternatives aiming to satisfy the criteria maximally. The generation process yields a range of suitable solutions. These solutions match the criteria outstandingly, while each one of them has different advantages and disadvantages. Exploring the trade‐offs among the solutions an architect is able to make more confident statements on the importance among criteria. This is because one can directly see the implications one's preferences have on the solutions. During our exploration among the solutions we do not need to bear the multitude of relations among design objects in mind. This is because they have already been taken into account in the computations. This way we are able to focus on the interplay among complex criteria rather than having to ensure the satisfaction of elemental requirements at the same time. This means we are able to exercise our preferences with great awareness. As a metaphor we can say we are conducting an orchestra focusing on the higher‐level aspects of the peace, in place of being busy to play every instrument at the same time. The approach is applied in two architectural design applications. One of the design tasks was devised from a real scene for computation; the other one was an actual project. In the latter application the computational results are compared against the solutions obtained using conventional means. The applications differ with respect to scale and context. They demonstrate the effectiveness and generic character of the approach. It is noted that possible applications of the method are diverse, due the generic character of the methodologies involved. That is, different requirements and intelligent objects can be formulated following the same approach. Therefore the cognitive approach is applicable throughout the various phases of the building process. These range from initiative to demolition phase, while the method remains essentially the same. Beyond architectural design the cognitive approach developed is relevant for decision making in complex problem domains, such as other areas of design and planning, including urban design, industrial design, and regional planning. It is emphasized that the approach is uniquely suitable when the problem includes multiple linguistic criteria, which is commonly the case in real‐world applications. Next to decision makers and researchers in these domains the thesis may also be of interest to researchers from the field of computational intelligence, who are interested in the particular problems of architecture, and the application potentials of their methodologies in this domain.
Benzer Tezler
- Intelligent design objects applied to the spatial allocation problem
Mekansal kurgu problemine akıllı tasarım objeleri uygulaması
JAVIER IGNACIO ZARATIEGUI FERNANDEZ
Yüksek Lisans
İngilizce
2014
MimarlıkOrta Doğu Teknik ÜniversitesiMimarlık Ana Bilim Dalı
DOÇ. DR. ARZU SORGUÇ
- Yapay zeka felsefesinde felsefi işlevselciliğin eleştirisi
The critique of the philosophical functionalism in the philosophy of artificial intelligence
TARIK TAZEGÜL
Yüksek Lisans
Türkçe
2023
Bilim ve Teknolojiİstanbul ÜniversitesiFelsefe Ana Bilim Dalı
DOÇ. DR. NECATİ MURAD OMAY
- William A. Dembski'nin akıllı tasarım kanıtı
William A. Dembski's intelligent design argument
AHMET ERKAN
Yüksek Lisans
Türkçe
2012
DinAnkara ÜniversitesiFelsefe ve Din Bilimleri Ana Bilim Dalı
PROF. DR. RECEP KILIÇ
- Uzman sistemler ve ulaştırma alanında kullanımları
Expert systems and using them in transportation
A.BURAK GÖKTEPE
- Yapay zeka uygulamaları
Artificial intelligence applications
PINAR AKKUŞ
Yüksek Lisans
Türkçe
1990
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrolİstanbul Teknik ÜniversitesiDOÇ.DR. MİTHAT UYSAL