Student Number : 9800233A - MSc dissertation - School of Electrical and Information Engineering - Faculty of Engineering and the Built Environment / Condition monitoring of machinery has increased in importance as more engineering processes are automated and the manpower required to operate and supervise plants is reduced. The monitoring of the condition of machinery can significantly reduce the cost of maintenance. Firstly, it can allow an early detection of potential catastrophic fault, which could be extremely expensive to repair. Secondly, it allows the implementation of conditions based maintenance rather than periodic or failure based maintenance [1]. In these cases, significant savings can be made by delaying schedule maintenance until convenient or necessary. Although there are numerous efficient methods for modeling of mechanical systems, they all suffer the disadvantage that they are only valid for a particular machine. Changes within the design or the operational mode of the machine normally require a manual adaptation. Using Neural Networks to model technical systems eliminates this major disadvantage. The basis for a successful model is an adequate knowledge base on which the network is "trained". Without prior knowledge of the machines systematic behavior or its history, training of a neural Network is not possible. Therefore, it is a pre-requisite that the knowledge base contains a complete behavior of the machine covering the respective operational modes whereby, not all rather the most important modes are required. Neural networks have a proven ability in the area of nonlinear pattern classification. After being trained, they contain expert knowledge and can correctly identify the different causes of bearing vibration. The capacity of artificial neural networks to mimic and automate human expertise is what makes them ideally suited for handling nonlinear systems. Neural networks are able to learn expert knowledge by being trained using a representative set of data [2]-[6]. At the beginning of a neural network’s training session, the neural network fault detector’s diagnosis of the motor’s condition will not be accurate. An error quantity is measured and used to adjust the neural network’s internal parameters in order to produce a more accurate output. This process is repeated until a suitable error is achieved. Once the network is sufficiently trained and the parameters have been saved, the neural network contains all the necessary knowledge to perform the fault detection. One of the most important aspects of achieving good neural network performance has proven to be the proper selection of training features. The curse of dimensionality states that, as a rule of thumb, the required cardinality of the training set for accurate training increases exponentially with the input dimension [7]. Thus feature selection which is a process of identifying those features that contribute most to the discrimination ability of the neural network is required. Proposed methods for selecting an appropriate subset of features are numerous [8]-[11]. Methods based on generating a single solution, such as the popular forward step wise approach, can fail to select features which do poorly alone but offer valuable information together. Approaches that maintain a population of solutions, such as genetic algorithms (GA) are more likely to speedily perform efficient searches in high dimensional spaces, with strong interdependencies among the features. The emphasis in using the genetic algorithm for feature selection is to reduce the computational load on the training system while still allowing near optimal results to be found relatively quickly. To obtain accurate measure of the condition of machinery, a wide range of approaches can be employed to select features indicative of condition. By comparing these features with features for known normal and probable fault conditions, the machine’s condition can be estimated. The most common approach is that of analysis in the frequency domain by applying a Fast Fourier Transform (FFT) to the time domain history data. The idea is simply to measure the energy (mean square value) of the vibrations. As the machine condition deteriorates, this measure is expected to increase. The method is able to reveal the harmonics around the fundamental frequency of the machine and other predominant frequency component (such as the cage frequency) [12]. Frequency analysis is well established and may be used to detect, diagnose and discriminate a variety of induction motor faults such as broken rotor bars, cage faults, phase imbalance, inner and outer race faults. However, as common in the monitoring of any industrial machine, background noise in recorded data can make spectra difficult to interpret. In addition, the accuracy of a spectrum is limited due to energy leakage [12- 14]. Like many of the new techniques now finding application in machinery condition monitoring, Higher Order Statistics was originally confined to the realms of non-linear structural dynamics. It has of recent however found successful application to the identification of abnormal operation of diesel engines and helicopter gearboxes [5, 7]. Higher Order Statistics provide convenient basis for comparison of data between different measurement instances and are sufficiently robust for on-line use. They are fast in computation compared with frequency or time-domain analysis. Furthermore, they give a more robust assessment than lower orders and can be used to calculate higher order spectra. This dissertation reports work which attempts to extend this capability to induction motors. The aim of this project is therefore to examine the use of Genetic Algorithms to select the most significant input features from a large set of possible features in machine condition monitoring contexts. The results show the effectiveness of the selected features from the acquired raw and preprocessed signals in diagnosis of machine condition. This project consists of the following tasks: #1; Using Fast Fourier transform and higher order signals techniques to preprocess data samples. #1; Create an intelligent engine using computational intelligence methods. The aim of this engine will be to recognize faulty bearings and assess the fault severity from sensor data. #1; Train the neural network using a back propagation algorithm. #1; Implement a feature selection algorithm using genetic algorithms to minimize the number of selected features and to maximize the performance of the neural network. #1; Retrain the neural network with the reduced set of features from genetic algorithm and compare the two approaches. #1; Investigate the effect of increasing the number of hidden nodes in the performance of the computational intelligence engine. #1; Evaluate the performance of the system using confusion matrices. The output of the design is the estimate of fault type and its severity, quantified on a scale between 0-3. Where, 0 corresponds to the absence of the specific fault and 3 the presence of a severe machine bearing fault. This research should make contribution to many sectors of industry such as electricity supply companies, and the railroad industry due to their need of techniques that are capable of accurately recognizing the development of a fault condition within a machine system component. Quality control of electric motors is an essential part of the manufacturing process as competition increases, the need for reliable and economical quality control becomes even more pressing. To this effect, this research project will contribute in the area of faults detection in the production line of electric motor.

A novel real-time control method that minimises linear system vibrations when it is subjected to an arbitrary external excitation is proposed in this study. The work deals with a discrete differential dynamic programming type of problem, in which an external disturbance is controlled over a time horizon by a control force strategy constituted by the well-known convolution approach. The proposed method states that if a control strategy can be established to restore an impulse external disturbance, then the convolution concept can be used to generate an overall control strategy to control the system response when it is subjected to an arbitrary external disturbance. The arbitrary disturbance is divided into impulses and by simply scaling, shifting and summation of the obtained control strategy against the impulse input for each impulse of the arbitrary disturbance, the overall control strategy will be established. Genetic Algorithm was adopted to obtain an optimal control force plan to suppress the system vibrations when it is subjected to a shock disturbance, and then the Convolution concept was used to enable the system response to be controlled in real-time using the obtained control strategy. Numerical tests were carried out on a two-degree of freedom quarter-vehicle active suspension model and the results were compared with results generated using the Linear Quadratic Regulator (LQR) method. The method was also applied to control the vibration of a seven-degree of freedom full-vehicle active suspension model. In addition, the effect of a time delay on the performance of the proposed approach was also studied. To demonstrate the applicability of the proposed method in real-time control, experimental tests were performed on a quarter-vehicle test rig equipped with a pneumatic active suspension. Numerical and experimental results showed the effectiveness of the proposed method in reducing the vehicle vibrations. One of the main contributions of this work besides using the Convolution concept to provide a real time control strategy is the reduction in the number of sensors needed to construct the proposed method as the disturbance amplitude is the only parameter needed to be measured (known). Finally, having achieved what has been proposed above, a generic robust control method is accomplished, which not only can be applied for active suspension systems but also in many other fields.

During times of stock market turbulence and crises, monitoring the clustering behaviour of financial instruments allows one to better understand the behaviour of the stock market and the associated systemic risks. In the study undertaken, I apply an effective and performant approach to classify data clusters in order to better understand correlations between stocks. The novel methods aim to address the lack of effective algorithms to deal with high-performance cluster analysis in the context of large complex real-time low-latency data-sets. I apply an efficient and novel data clustering approach, namely the Giada and Marsili log-likelihood function derived from the Noh model and use a Parallel Genetic Algorithm in order to isolate residual data clusters. Genetic Algorithms (GAs) are a very versatile methodology for scientific computing, while the application of Parallel Genetic Algorithms (PGAs) further increases the computational efficiency. They are an effective vehicle to mine data sets for information and traits. However, the traditional parallel computing environment can be expensive. I focused on adopting NVIDIAs Compute Unified Device Architecture (CUDA) programming model in order to develop a PGA framework for my computation solution, where I aim to efficiently filter out residual clusters. The results show that the application of the PGA with the novel clustering function on the CUDA platform is quite effective to improve the computational efficiency of parallel data cluster analysis.

by Lee Wing Kin. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2000. / Includes bibliographical references (leaves 81-90). / Abstracts in English and Chinese. / Abstract --- p.i / Acknowledgments --- p.iii / List of Figures --- p.vii / List of Tables --- p.viii / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Clustering --- p.1 / Chapter 1.1.1 --- Hierarchical Classification --- p.2 / Chapter 1.1.2 --- Partitional Classification --- p.3 / Chapter 1.1.3 --- Comparative Analysis --- p.4 / Chapter 1.2 --- Cluster Analysis and Traveling Salesman Problem --- p.5 / Chapter 1.3 --- Solving Clustering Problem --- p.7 / Chapter 1.4 --- Genetic Algorithms --- p.9 / Chapter 1.5 --- Outline of Work --- p.11 / Chapter 2 --- The Clustering Algorithms and Applications --- p.13 / Chapter 2.1 --- Introduction --- p.13 / Chapter 2.2 --- Traveling Salesman Problem --- p.14 / Chapter 2.2.1 --- Related Work on TSP --- p.14 / Chapter 2.2.2 --- Solving TSP using Genetic Algorithm --- p.15 / Chapter 2.3 --- Applications --- p.22 / Chapter 2.3.1 --- Clustering for Vertical Partitioning Design --- p.22 / Chapter 2.3.2 --- Horizontal Partitioning a Relational Database --- p.36 / Chapter 2.3.3 --- Object-Oriented Database Design --- p.42 / Chapter 2.3.4 --- Document Database Design --- p.49 / Chapter 2.4 --- Conclusions --- p.53 / Chapter 3 --- The Experiments for Vertical Partitioning Problem --- p.55 / Chapter 3.1 --- Introduction --- p.55 / Chapter 3.2 --- Comparative Study --- p.56 / Chapter 3.3 --- Experimental Results --- p.59 / Chapter 3.4 --- Conclusions --- p.61 / Chapter 4 --- Three New Operators for TSP --- p.62 / Chapter 4.1 --- Introduction --- p.62 / Chapter 4.2 --- Enhanced Cost Edge Recombination Operator --- p.63 / Chapter 4.3 --- Shortest Path Operator --- p.66 / Chapter 4.4 --- Shortest Edge Operator --- p.69 / Chapter 4.5 --- The Experiments --- p.71 / Chapter 4.5.1 --- Experimental Results for a 48-city TSP --- p.71 / Chapter 4.5.2 --- Experimental Results for Problems in TSPLIB --- p.73 / Chapter 4.6 --- Conclusions --- p.77 / Chapter 5 --- Conclusions --- p.78 / Chapter 5.1 --- Summary of Achievements --- p.78 / Chapter 5.2 --- Future Development --- p.80 / Bibliography --- p.81

De nombreuses fonctions cellulaires sont réalisées grâce à l'interaction coordonnée de plusieurs gènes. Identifier le graphe de ces interactions, appelé réseau de régulation génétique, à partir de données d'expression de gènes est l'un des objectifs majeurs de la biologie des systèmes. Dans cette thèse, nous abordons ce problème en choisissant de modéliser les relations entre gènes par un réseau bayésien. Se pose alors la question de l'apprentissage de la structure de ce type de modèle à partir de données qui sont en général peu nombreuses. Pour résoudre ce problème, nous recherchons parmi tous les modèles possibles le modèle le plus simple, expliquant le mieux les données. Pour cela, nous introduisons et étudions différents types d'algorithmes génétiques permettant d'explorer l'espace des modèles. Nous nous intéressons plus particulièrement aux méthodes de spéciation. ces dernières, en favorisant la diversité des solutions candidates considérées, empêchent l'algorithme de converger trop rapidement vers des optima locaux. Ces algorithmes génétiques sont comparés avec différentes méthodes d'apprentissage de structure de réseaux bayésiens, classiquement utilisées dans la littérature. Nous mettons ainsi en avant la pertinence des approches evolutionnaires pour l'apprentissage de ces graphes d'interactions. Enfin, nous les comparons à une classe alternative d'algorithmes évolutionnaires qui s'avère particulièrement prometteuse : les algorithmes à estimation de distribution. Tous ces algorithmes sont testés et comparés sur un modèle du réseau de régulation de l'insuline de 35 noeuds dont nous tirons des jeux de données synthétiques de taille modeste. / Inferring gene regulatory networks from data requires the development of algorithms devoted to structure extraction. When only static data are available, gene interactions may be modelled by a bayesian network that represents the presence of direct interactions from regulators to regulees by conditional probability distributions. In this work, we used enhanced evolutionary algorithms to stochastically evolve a set of candidate bayesian network structures and found the model that best fits data without prior knowledge. We proposed various evolutionary strategies suitable for the task and tested our choices using simulated data drawn from a given bio-realistic network of 35 nodes, the so-called insulin network, which has been used in the literature for benchmarking. We introduced a niching strategy that reinforces diversity through the population and avoided trapping of the algorithm in one local minimum in the early steps of learning. We compared our best evolutionary approach with various well known learning algorithms (mcmc, k2, greedy search, tpda, mmhc) devoted to bayesian network structure learning. Then, we compared our best genetic algorithm with another class of evolutionary algorithms : estimation of distribution algorithms. We show that an evolutionary approach enhanced by niching outperforms classical structure learning methods in elucidating the original model. Finally, it appears that estimation of distribution algorithms are a promising approach to extend this work. These results were obtained for the learning of a bio-realistic network and, more importantly, on various small datasets.

Lee Sum Wai = 應用多目標基因演算法於合併三維破裂物件 / 李芯慧. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2004. / Includes bibliographical references. / Text in English; abstracts in English and Chinese. / Lee Sum Wai = Ying yong duo mu biao ji yin yan suan fa yu he bing san wei po lie wu jian / Li Xinhui. / Contents --- p.VI / List of Figures --- p.IX / List of Tables --- p.XIII / Chapter Chapter 1 --- Introduction --- p.1-1 / Chapter 1.1. --- A review of assembling objects --- p.1-3 / Chapter 1.1.1. --- Two-Dimensional matching --- p.1-3 / Chapter 1.1.2. --- Three-Dimensional matching --- p.1-4 / Chapter 1.1.3. --- 2.5-Dimensional matching --- p.1-5 / Chapter 1.2. --- Objectives of this research work --- p.1-7 / Chapter 1.2.1. --- Local Matching of fragments --- p.1-7 / Chapter 1.2.2. --- Global Matching fragments --- p.1-8 / Chapter 1.3. --- Thesis Outline --- p.1-9 / Chapter Chapter 2 --- Background Information --- p.2-1 / Chapter 2.1. --- Three-Dimensional Objects Representation --- p.2-1 / Chapter 2.2. --- Three-Dimensional Objects Geometric Transformation --- p.2-3 / Chapter 2.1.1. --- Translation --- p.2-4 / Chapter 2.1.2. --- Rotation --- p.2-5 / Chapter 2.3. --- Orientated Bounding Box (OBB) --- p.2-6 / Chapter 2.4. --- Scan-Line Method --- p.2-7 / Chapter 2.5. --- Mesh Simplification --- p.2-10 / Chapter 2.6. --- Review of the Surface Matching Method --- p.2-12 / Chapter 2.6.1. --- G. Papaioannou et al ´بs method --- p.2-13 / Chapter Chapter 3 --- Genetic Algorithm --- p.3-1 / General introduction --- p.3-1 / Chapter 3.1. --- Characteristics of Genetic Algorithms --- p.3-3 / Chapter 3.2. --- Mechanism of Genetic Algorithms --- p.3-4 / Chapter 3.2.1. --- Coding --- p.3-4 / Chapter 3.2.2. --- Reproduction --- p.3-5 / Chapter 3.2.3. --- Selection --- p.3-8 / Chapter 3.2.4. --- Stopping Criteria --- p.3-9 / Chapter 3.3. --- Convergence of Genetic Algorithms --- p.3-10 / Chapter 3.4. --- Comparison with Traditional Optimization Methods --- p.3-13 / Chapter 3.4.1. --- Test Function - Sphere --- p.3-14 / Chapter 3.4.2. --- Test Function - Rosenbrock's Saddle --- p.3-19 / Chapter 3.4.3. --- Test Function 一 Step --- p.3-22 / Chapter 3.4.4. --- Test Function -Quartic --- p.3-25 / Chapter 3.4.5. --- Test Function - Shekel's Foxholes --- p.3-28 / Chapter 3.5. --- Multi-Objective Genetic Algorithms --- p.3-29 / Chapter 3.5.1. --- Non-Pareto Approach --- p.3-31 / Chapter 3.5.2. --- Pareto-Ranking --- p.3-32 / Chapter 3.5.3. --- Comparison --- p.3-35 / Chapter Chapter 4 --- Assembling broken objects (I) --- p.4-1 / Chapter 4.1. --- System Flow of Single Pair Assemblage --- p.4-2 / Chapter 4.2. --- Parameterization --- p.4-3 / Chapter 4.2.1. --- Degree of Freedom --- p.4-3 / Chapter 4.2.2. --- Reference Plane and Sampling Points --- p.4-4 / Chapter 4.3. --- Matching Error --- p.4-5 / Chapter 4.3.1. --- Counterpart Surface Matching Error --- p.4-5 / Chapter 4.3.2. --- Border Matching Error --- p.4-7 / Chapter 4.4. --- Correlation-Based Matching Method --- p.4-14 / Chapter Chapter 5 --- Assembling Broken Objects (II)- Global Matching --- p.5-1 / Chapter 5.1. --- Arrangement Strategy --- p.5-2 / Chapter 5.1.1. --- Introduction to Packing --- p.5-2 / Chapter 5.1.2. --- Proposed Architecture --- p.5-6 / Chapter 5.2. --- Relational Multi-Objective Genetic Algorithm --- p.5-13 / Chapter 5.2.1. --- Existing Problem --- p.5-13 / Chapter 5.2.2. --- A New Operator --- p.5-14 / Chapter 5.2.3. --- Relationship Function --- p.5-16 / Chapter 5.3. --- Conclusion and summary --- p.5-20 / Chapter Chapter 6 --- Optimization Approach by Genetic Algorithm --- p.6-1 / Chapter 6.1. --- Solution Space --- p.6-1 / Chapter 6.2. --- Formulation of Gene and Chromosome --- p.6-3 / Chapter 6.2.1. --- Matching Three or More Fragments --- p.6-4 / Chapter 6.2.2. --- Matching Two Fragments --- p.6-5 / Chapter 6.3. --- Fitness Function --- p.6-5 / Chapter 6.3.1. --- Matching Two Fragments --- p.6-5 / Chapter 6.3.2. --- Matching Three or More Fragments --- p.6-6 / Chapter 6.4. --- Reproduction --- p.6-7 / Chapter 6.4.1. --- Crossover --- p.6-8 / Chapter 6.4.2. --- Mutation --- p.6-9 / Chapter 6.4.3. --- Inheritance --- p.6-9 / Chapter 6.5. --- Selection --- p.6-9 / Chapter Chapter 7 --- Experimental Results --- p.7-1 / Chapter 7.1 --- Data Acquisition --- p.7-1 / Chapter 7.2 --- Experiment for Mesh Simplification --- p.7-4 / Chapter 7.3 --- Experiment for Correlation-Based Matching Method --- p.7-5 / Chapter 7.4 --- Experiment One: Two Fragments --- p.7-6 / Chapter 7.5 --- Experiment Two: Several Fragments --- p.7-10 / Chapter 7.5.1 --- Constraint Direction Matching --- p.7-10 / Chapter 7.5.2 --- Unconstraint Direction Matching --- p.7-14 / Chapter Chapter 8 --- Conclusion --- p.8-1 / Appendix Reference --- p.1

Ng Sai-Cheong. / Thesis submitted in: December 2004. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2005. / Includes bibliographical references (leaves 172-178). / Abstracts in English and Chinese. / Abstract --- p.i / Acknowledgement --- p.iii / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Data Mining --- p.1 / Chapter 1.2 --- Problem Specifications and Motivations --- p.3 / Chapter 1.3 --- Contributions of the Thesis --- p.5 / Chapter 1.4 --- Thesis Roadmap --- p.6 / Chapter 2 --- Related Work --- p.9 / Chapter 2.1 --- Supervised Classification Techniques --- p.9 / Chapter 2.1.1 --- Classification Rules --- p.9 / Chapter 2.1.2 --- Decision Trees --- p.11 / Chapter 2.2 --- Evolutionary Algorithms --- p.19 / Chapter 2.2.1 --- Genetic Algorithms --- p.19 / Chapter 2.2.2 --- Genetic Programming --- p.24 / Chapter 2.2.3 --- Evolution Strategies --- p.26 / Chapter 2.2.4 --- Evolutionary Programming --- p.32 / Chapter 2.3 --- Applications of Evolutionary Algorithms to Induction of Classification Rules --- p.33 / Chapter 2.3.1 --- SCION --- p.33 / Chapter 2.3.2 --- GABIL --- p.34 / Chapter 2.3.3 --- LOGENPRO --- p.35 / Chapter 2.4 --- Applications of Evolutionary Algorithms to Construction of Decision Trees --- p.35 / Chapter 2.4.1 --- Binary Tree Genetic Algorithm --- p.35 / Chapter 2.4.2 --- OC1-GA --- p.36 / Chapter 2.4.3 --- OC1-ES --- p.38 / Chapter 2.4.4 --- GATree --- p.38 / Chapter 2.4.5 --- Induction of Linear Decision Trees using Strong Typing GP --- p.39 / Chapter 2.5 --- Spatial Data Structures and its Applications --- p.40 / Chapter 2.5.1 --- Spatial Data Structures --- p.40 / Chapter 2.5.2 --- Applications of Spatial Data Structures --- p.42 / Chapter 3 --- Induction of Classification Rules using Genetic Algorithms --- p.45 / Chapter 3.1 --- Introduction --- p.45 / Chapter 3.2 --- Rule Learning using Genetic Algorithms --- p.46 / Chapter 3.2.1 --- Population Initialization --- p.47 / Chapter 3.2.2 --- Fitness Evaluation of Chromosomes --- p.49 / Chapter 3.2.3 --- Token Competition --- p.50 / Chapter 3.2.4 --- Chromosome Elimination --- p.51 / Chapter 3.2.5 --- Rule Migration --- p.52 / Chapter 3.2.6 --- Crossover --- p.53 / Chapter 3.2.7 --- Mutation --- p.55 / Chapter 3.2.8 --- Calculating the Number of Correctly Classified Training Samples in a Rule Set --- p.56 / Chapter 3.3 --- Performance Evaluation --- p.56 / Chapter 3.3.1 --- Performance Comparison of the GA-based CPRLS and Various Supervised Classifi- cation Algorithms --- p.57 / Chapter 3.3.2 --- Performance Comparison of the GA-based CPRLS and RS-based CPRLS --- p.68 / Chapter 3.3.3 --- Effects of Token Competition --- p.69 / Chapter 3.3.4 --- Effects of Rule Migration --- p.70 / Chapter 3.4 --- Chapter Summary --- p.73 / Chapter 4 --- Genetic Algorithm-based Quadratic Decision Trees --- p.74 / Chapter 4.1 --- Introduction --- p.74 / Chapter 4.2 --- Construction of Quadratic Decision Trees --- p.76 / Chapter 4.3 --- Evolving the Optimal Quadratic Hypersurface using Genetic Algorithms --- p.77 / Chapter 4.3.1 --- Population Initialization --- p.80 / Chapter 4.3.2 --- Fitness Evaluation --- p.81 / Chapter 4.3.3 --- Selection --- p.81 / Chapter 4.3.4 --- Crossover --- p.82 / Chapter 4.3.5 --- Mutation --- p.83 / Chapter 4.4 --- Performance Evaluation --- p.84 / Chapter 4.4.1 --- Performance Comparison of the GA-based QDT and Various Supervised Classification Algorithms --- p.85 / Chapter 4.4.2 --- Performance Comparison of the GA-based QDT and RS-based QDT --- p.92 / Chapter 4.4.3 --- Effects of Changing Parameters of the GA-based QDT --- p.93 / Chapter 4.5 --- Chapter Summary --- p.109 / Chapter 5 --- Induction of Linear and Quadratic Decision Trees using Spatial Data Structures --- p.111 / Chapter 5.1 --- Introduction --- p.111 / Chapter 5.2 --- Construction of k-D Trees --- p.113 / Chapter 5.3 --- Construction of Generalized Quadtrees --- p.119 / Chapter 5.4 --- Induction of Oblique Decision Trees using Spatial Data Structures --- p.124 / Chapter 5.5. --- Induction of Quadratic Decision Trees using Spatial Data Structures --- p.130 / Chapter 5.6 --- Performance Evaluation --- p.139 / Chapter 5.6.1 --- Performance Comparison with Various Supervised Classification Algorithms --- p.142 / Chapter 5.6.2 --- Effects of Changing the Minimum Number of Training Samples at Each Node of a k-D Tree --- p.155 / Chapter 5.6.3 --- Effects of Changing the Minimum Number of Training Samples at Each Node of a Generalized Quadtree --- p.157 / Chapter 5.6.4 --- Effects of Changing the Size of Datasets . --- p.158 / Chapter 5.7 --- Chapter Summary --- p.160 / Chapter 6 --- Conclusions --- p.164 / Chapter 6.1 --- Contributions --- p.164 / Chapter 6.2 --- Future Work --- p.167 / Chapter A --- Implementation of Data Mining Algorithms Specified in the Thesis --- p.170 / Bibliography --- p.178

by Yong Liang. / "November 2003." / Thesis (Ph.D.)--Chinese University of Hong Kong, 2003. / Includes bibliographical references (p. 237-266). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. Ann Arbor, MI : ProQuest Information and Learning Company, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Mode of access: World Wide Web. / Abstracts in English and Chinese.

Lee Wai Hung. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1996. / Includes bibliographical references (leaves 108-115). / Chapter 1 --- Introduction --- p.8 / Chapter 1.1 --- Introduction to Group Technology --- p.8 / Chapter 1.2 --- Cell design --- p.9 / Chapter 1.3 --- Objectives of the research --- p.11 / Chapter 1.4 --- Organization of thesis --- p.11 / Chapter 2 --- Literature review --- p.13 / Chapter 2.1 --- Introduction --- p.13 / Chapter 2.2 --- Standard models --- p.14 / Chapter 2.2.1 --- Array-based methods --- p.16 / Chapter 2.2.2 --- Cluster identification --- p.16 / Chapter 2.2.3 --- Graph-based methods --- p.17 / Chapter 2.2.4 --- Integer programming --- p.17 / Chapter 2.2.5 --- Seed-based --- p.18 / Chapter 2.2.6 --- Similarity coefficient --- p.18 / Chapter 2.2.7 --- Artificial intelligence methods --- p.19 / Chapter 2.3 --- Generalized models --- p.19 / Chapter 2.3.1 --- Machine assignment models --- p.20 / Chapter 2.3.2 --- Part family models --- p.20 / Chapter 2.3.3 --- Cell formation models --- p.21 / Chapter 3 --- Genetic cell formation algorithm --- p.22 / Chapter 3.1 --- Introduction --- p.22 / Chapter 3.2 --- TSP formulation for a permutation of machines --- p.23 / Chapter 3.3 --- Genetic algorithms --- p.26 / Chapter 3.3.1 --- Representation and basic crossover operators --- p.27 / Chapter 3.3.2 --- Fitness function --- p.28 / Chapter 3.3.3 --- Initialization --- p.29 / Chapter 3.3.4 --- Parent selection strategies --- p.30 / Chapter 3.3.5 --- Crossover --- p.31 / Chapter 3.3.6 --- Mutation --- p.37 / Chapter 3.3.7 --- Replacement --- p.38 / Chapter 3.3.8 --- Termination --- p.38 / Chapter 3.4 --- Formation of machine cells and part families --- p.39 / Chapter 3.4.1 --- Objective functions --- p.39 / Chapter 3.4.2 --- Machine assignment --- p.42 / Chapter 3.4.3 --- Part assignment --- p.43 / Chapter 3.5 --- Implementation --- p.43 / Chapter 3.6 --- An illustrative example --- p.45 / Chapter 3.7 --- Comparative Study --- p.49 / Chapter 3.8 --- Conclusions --- p.50 / Chapter 4 --- A multi-chromosome GA for minimizing total intercell and intracell moves --- p.55 / Chapter 4.1 --- Introduction --- p.55 / Chapter 4.2 --- The model --- p.57 / Chapter 4.3 --- Solution techniques to the workload model --- p.61 / Chapter 4.3.1 --- Logendran's original approach --- p.62 / Chapter 4.3.2 --- Standard representation - the GA approach --- p.63 / Chapter 4.3.3 --- Multi-chromosome representation --- p.65 / Chapter 4.4 --- Comparative Study --- p.70 / Chapter 4.4.1 --- Problem 1 --- p.70 / Chapter 4.4.2 --- Problem 2 --- p.71 / Chapter 4.4.3 --- Problem 3 --- p.75 / Chapter 4.4.4 --- Problem 4 --- p.76 / Chapter 4.5 --- Bi-criteria Model --- p.79 / Chapter 4.5.1 --- Experimental results --- p.85 / Chapter 4.6 --- Conclusions --- p.85 / Chapter 5 --- Integrated design of cellular manufacturing systems in the presence of alternative process plans --- p.88 / Chapter 5.1 --- Introduction --- p.88 / Chapter 5.1.1 --- Literature review --- p.90 / Chapter 5.1.2 --- Motivation --- p.92 / Chapter 5.2 --- Mathematical models --- p.93 / Chapter 5.2.1 --- Notation --- p.93 / Chapter 5.2.2 --- Objective functions --- p.95 / Chapter 5.3 --- Our solution --- p.96 / Chapter 5.4 --- Illustrative example and analysis of results --- p.98 / Chapter 5.4.1 --- Solution for objective function 1 --- p.101 / Chapter 5.4.2 --- Solution for objective function 2 --- p.102 / Chapter 5.5 --- Conclusions --- p.103 / Chapter 6 --- Conclusions --- p.104 / Chapter 6.1 --- Summary of achievements --- p.104 / Chapter 6.2 --- Future works --- p.106

I develop the Generalized Evolutionary Nash Equilibrium Estimator (GENEE) library. The tool is designed to provide a generic computational library for running genetic algorithms and individual evolutionary learning in economic decision-making environments. Most importantly, I have adapted the library to estimate equilibria bidding functions in auctions. I show it produces highly accurate estimates across a large class of auction environments with known solutions. I then apply GENEE to estimate the equilibria of two additional auctions with no known solutions: first-price sealed-bid common value auctions with multiple signals, and simultaneous first-price auctions with subadditive values