Thesis (M.S.)--Missouri University of Science and Technology, 2008. / Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed April 26, 2008) Includes bibliographical references (p. 65-68).

Thesis (M.S.)--Missouri University of Science and Technology, 2008. / Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed April 29, 2008) Includes bibliographical references (p. 53-55).

Thesis (M.S.)--Missouri University of Science and Technology, 2009. / Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed August 10, 2009) Includes bibliographical references (p. 49-51).

Thesis (M.S.)--Missouri University of Science and Technology, 2010. / Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed July 19, 2010) Includes bibliographical references (p. 56-62).

- / -

学位授与大学：京都大学 ; 取得学位: 博士(工学) ; 学位授与年月日: 2007-09-25 ; 学位の種類: 新制・課程博士 ; 学位記番号: 工博第2855号 ; 請求記号: 新制/工/1420 ; 整理番号: 25540 / Design problem solving behavior refers to the way in which people solve their creative problem of design in their mind. It is one of the basic problems in the area of design methodology, which varies greatly by cases and designers. On the other hand, there are still some general ways or commonness as the core. Because of the complexity of design problem solving behavior, it is still not understood very well. This dissertation dives into the problem of design problem solving behavior too and tried to provide a general view of it, including both the general strategies and the temporary tactics. But differs from many other researches, it employed a confined and well-structured simulation of manual design process by employing the method of interactive evolutionary computation (IEC) to extract design problem solving behavior objectively. The simulated design process provided a comparable and statistically analyzable model for exploring design problem solving behavior of people, and made the findings of this dissertation more reliable. The design problem of interior works of Chinese residents, which need little special knowledge to solve, was selected as the design problem in this dissertation. The method of IEC was applied in interior works design for helping the Chinese residents to solve the practical interior works design problems, and inducing the design problem solving behavior of them. The dissertation contains 6 chapters, including the general introduction (chapter 1), the main body (chapter 2 to 5), and the conclusion (chapter 6). The main body can be further divided into two parts. In the first part (chapter 2 and 3) the IEC interior works (IECIW) design system was developed, and evaluated by a large amount of Chinese residents on its usability and disadvantage. After the preparation of method in the first part, the second part (chapter 4 and 5) presented two parallel researches on participants’ design problem solving behavior in design process using IEC in order to approach the design problem solving behavior in common design processes. Chapter 1 introduces the background and purpose of the research, reviewed related literatures, and the frame work of the dissertation. In chapter 2, IEC method was tentatively applied in the problem of interior works design. 7 color and texture related factors of the living room of a typical apartment in Beijing were selected as design factors in the IEC IW design system. Through 3 experiments, the IEC IW design system was found effective in interior works design and heuristic for the two tested Chinese students. The effect of increasing population size was also found significantly increasing the efficiency of the system. In chapter 3, the developed IEC IW design system was tentatively used by 231 Chinese residents to evaluate its usability and disadvantage in real design problems of interior works. It was concluded that the IEC IW design system is useful for the residents, and it was also found that older participants, and those with lower education and family income levels, gave the system better evaluations. Chapter 4 started to explore problem solving behavior of people in design tasks through simulated design process for interior works using IEC. Data of design process employing IEC of 8 Chinese participants were collected. Through analysis of design problem solving process, it was revealed that people tend to do what they are certain of firstly, and make harder decisions later. It was also found that people did not tend to move their eyes to a faraway image in the interface constantly, which was considered more convenient for them. Chapter 5 continued to explore problem solving behavior of the 8 participants' interior works design process employing IEC. The method of protocol analysis was employed to analyze verbal reports of the participants. It was revealed that different parts of the interior scene have different influence on people's evaluation, and people tended to use same evaluation criterion continuously on several images, then switch to another evaluation criterion. 3 stages of design problem solving behavior along the process were also explained. Chapter 6 summarizes the findings in the dissertation, presents the general discussion and perspective, and proposed some research in the future. / Kyoto University (京都大学) / 0048 / 新制・課程博士 / 博士(工学) / 甲第13384号 / 工博第2855号 / 新制||工||1420(附属図書館) / 25540 / UT51-2007-Q785 / 京都大学大学院工学研究科建築学専攻 / (主査)教授 宗本 順三, 教授 上谷 宏二, 教授 加藤 直樹 / 学位規則第4条第1項該当

This thesis describes the investigations undertaken to produce a novel hybrid optimisation technique that combines both global and local searching to produce good solutions quickly. Many evolutionary computation and experimental design methods are considered before genetic algorithms and evolutionary operation are combined to produce novel optimisation algorithms. A novel piece of software is created to run two and three factor evolutionary operation experiments. A range of new hybrid small population genetic algorithms are created that contain evolutionary operation in all generations (static hybrids) or contain evolutionary operation in a controlled number of generations (dynamic hybrids). A large number of empirical tests are carried out to determine the influence of operators and the performance of the hybrids over a range of standard test functions. For very small populations, twenty or less individuals, stochastic universal sampling is demonstrated to be the most suitable method of selection. The performance of very small population evolutionary operation hybrid genetic algorithms is shown to improve with larger generation gaps on simple functions and on more complex functions increasing the generation gap does not deteriorate performance. As a result of the testing carried out for this study a generation gap of 0.7 is recommended as a starting point for empirical searches using small population genetic algorithms and their hybrids. Due to the changing presence of evolutionary operation, the generation gap has less influence on dynamic hybrids compared to the static hybrids. The evolutionary operation, local search element is shown to positively influence the performance of the small population genetic algorithm search. The evolutionary operation element in the hybrid genetic algorithm gives the greatest improvement in performance when present in the middle generations or with a progressively greater presence. A recommendation for the information required to be reported for benchmarking genetic algorithm performance is also presented. This includes processor, platform, software information as well as genetic algorithm parameters such as population size, number of generations, crossover method and selection operators and results of testing on a set of standard test functions.

This work presents the design, evaluation and comparison of evolution and reinforcement learning models, in isolation and combined in Darwinian and Lamarckian frameworks, with a particular emphasis being placed on their adaptive nature in response to environments that become increasingly unstable. Our ultimate objective is to determine whether hybrid models of evolution and learning can demonstrate adaptive qualities beyond those of such models when applied in isolation. This work demonstrates the limitations of evolution, reinforcement learning and Lamarckian models in dealing with increasingly unstable environments, while noting the effective adaptive nature of a Darwinian model to assimilate increasing levels of instability. This is shown to be a result of the Darwinian evolution model's ability to separate learning at two levels, the population's experience of the environment over the course of many generations and the individual's experience of the environment over the course of its lifetime. Thus, knowledge relating to the general characteristics of the environment over many generations can be maintained in the population's genotypes with phenotype (reinforcement) learning being utilized to adapt a particular agent to the particular characteristics of its environment. Lamarckian evolution, though, is shown to demonstrate adaptive characteristics that are highly effective in response to the stable environments. Selection and reproduction combined with reinforcement learning creates a model that has the ability to utilize useful knowledge produced by reinforcements, as opposed to random mutations, to accelerate the search process. As a result the influence of individual learning on the populations evolution is shown to be more successful when applied in the more direct Lamarckian form. Based on our results demonstrating the success of Lamarckian strategies in stable environments and Darwinian strategies in unstable environments, hybrid Darwinian/Lamarckian models are created with a view towards combining the advantages of both forms of evolution to produce a superior adaptive capability. Our investigation demonstrates that such hybrid models can effectively combine the adaptive advantageous of both Darwinian and Lamarckian evolution to provide a more effective capability of adapting to a range of conditions, from stable to unstable, appropriately adjusting the required degree of inheritance in response to the requirements of the environment.

The impulse shared by all humans to express ourselves through dance represents a unique opportunity to artificially capture human creative expression. 1hls ambition aligns with the aim of artificial intelligence (AI) to study and emulate those aspects of human intelligence that are not readily reproduced in existing computer algorithms. As a first step toward addressing this challenge, this thesis describes Dance Evolution, which focuses on movements that are tied to a specific beat of music. Furthermore, Dance Evolution harnesses the users own taste to ex pl ore the new and interesting dances, allowing ta novel form of self-expression mediated by the computer, following the trend started by music and rhythm games. By implementing an algorithm that identifies the most prominent sounds within a song, Dance Evolution in effect allows artificial neural networks (ANNs) to listen to any song and exploit its rhythmic structure. Interactive evolution provides a tool for users to search increasingly intricate movement sequences by breeding their ANN controllers, in the same way that a gardener might explore interesting plants by breeding hybrids. The underlying idea in Dance Evolution is thus to create a novel mapping between sound and movement that evokes the spirit of casually dancing to the beat of a song.

The evolution of explicitly represented topologies such as graphs involves devising methods for mutating, comparing and combining structures in meaningful ways and identifying and maintaining the necessary topological diversity. Research has been conducted in the area of the evolution of trees in genetic programming and of neural networks and some of these problems have been addressed independently by the different research communities. In the domain of neural networks, NEAT (Neuroevolution of Augmenting Topologies) has shown to be a successful method for evolving increasingly complex networks. This system's success is based on three interrelated elements: speciation, marking of historical information in topologies, and initializing search in a small structures search space. This provides the dynamics necessary for the exploration of diverse solution spaces at once and a way to discriminate between different structures. Although different representations have emerged in the area of genetic programming, the study of the tree representation has remained of interest in great part because of its mapping to programming languages and also because of the observed phenomenon of unnecessary code growth or bloat which hinders performance. The structural similarity between trees and neural networks poses an interesting question: Is it possible to apply the techniques from NEAT to the evolution of trees and if so, how does it affect performance and the dynamics of code growth? In this work we address these questions and present analogous techniques to those in NEAT for genetic programming.