"After decades of development, Intelligent Tutoring Systems (ITSs) have become a common learning environment for learners of various domains and academic levels. ITSs are computer systems designed to provide instruction and immediate feedback, which is customized to individual students, but without requiring the intervention of human instructors. All ITSs share the same goal: to provide tutorial services that support learning. Since learning is a very complex process, it is not surprising that a range of technologies and methodologies from different fields is employed. Student modeling is a pivotal technique used in ITSs. The model observes student behaviors in the tutor and creates a quantitative representation of student properties of interest necessary to customize instruction, to respond effectively, to engage students¡¯ interest and to promote learning. In this dissertation work, I focus on the following aspects of student modeling. Part I: Student Knowledge: Parameter Interpretation. Student modeling is widely used to obtain scientific insights about how people learn. Student models typically produce semantically meaningful parameter estimates, such as how quickly students learn a skill on average. Therefore, parameter estimates being interpretable and plausible is fundamental. My work includes automatically generating data-suggested Dirichlet priors for the Bayesian Knowledge Tracing model, in order to obtain more plausible parameter estimates. I also proposed, implemented, and evaluated an approach to generate multiple Dirichlet priors to improve parameter plausibility, accommodating the assumption that there are subsets of skills which students learn similarly. Part II: Student Performance: Student Performance Prediction. Accurately predicting student performance is one of the most desired features common evaluations for student modeling. for an ITS. The task, however, is very challenging, particularly in predicting a student¡¯s response on an individual problem in the tutor. I analyzed the components of two common student models to determine which aspects provide predictive power in classifying student performance. I found that modeling the student¡¯s overall knowledge led to improved predictive accuracy. I also presented an approach, which, rather than assuming students are drawn from a single distribution, modeled multiple distributions of student performances to improve the model¡¯s accuracy. Part III: Wheel-spinning: Student Future Failure in Mastery Learning. One drawback of the mastery learning framework is its possibility to leave a student stuck attempting to learn a skill he is unable to master. We refer to this phenomenon of students being given practice with no improvement as wheel-spinning. I analyzed student wheel-spinning across different tutoring systems and estimated the scope of the problem. To investigate the negative consequences of see what wheel-spinning could have done to students, I investigated the relationships between wheel-spinning and two other constructs of interest about students: efficiency of learning and ¡°gaming the system¡±. In addition, I designed a generic model of wheel-spinning, which uses features easily obtained by most ITSs. The model can be well generalized to unknown students with high accuracy classifying mastery and wheel-spinning problems. When used as a detector, the model can detect wheel-spinning in its early stage with satisfying satisfactory precision and recall. "

ASSISTments is a web-based intelligent tutoring system that can provide students with immediate feedback when they are doing math homework. Until now, ASSISTments required internet access in order to do nightly homework. Without ASSISTments, students do their work on paper and are not told if they are correct or given help for wrong answers until the next morning at best. We've developed a component that supports 'offline-mode', enabling students without internet access at home to still receive immediate feedback on their responses. Students with laptops download their assignments at school, and then run ASSISTments at home in offline mode, utilizing the browser's application cache and Web Storage API. To evaluate the benefit of having the offline feature, we ran a randomized controlled study that tests the effect of immediate feedback on student learning. Intuition would suggest that providing a student with tutoring and feedback immediately after they submit an answer would lead to better understanding of the material than having them wait until the next day. The results of the study confirmed our hypothesis, and validated the need for 'offline mode.'

This study explores the ability of an Intelligent Tutoring System (ITS) to increase parental engagement in student learning. A parental notification feature was developed for the web-based ASSISTments ITS that allows parents to log into their own accounts and access detailed data about their students' performance. Parents from a local middle school were then invited to create accounts and answer a survey assessing how engaged they felt they were in their students' education. A randomized controlled experiment was run during which weekly automated messages were sent home to parents regarding their students' assignments and how they were performing. After having them take a post-survey, it was found that access to this data caused parents to become more involved in their students' education. Additionally, this led to increased student performance in the form of higher homework completion rates. Qualitative feedback from parents was very positive.

This research explores the possibility of reproducing mixing decisions of a skilled audio engineer with minimal human interaction that can improve the overall listening experience of musical mixtures, i.e., intelligent mixing. By producing a balanced mix automatically musician and mixing engineering can focus on their creativity while the productivity of music production is increased. We focus on the two essential aspects of such a system, frequency and dynamics. This thesis presents an intelligent strategy for multitrack frequency and dynamics processing that exploit the interdependence of input audio features, incorporates best practices in audio engineering, and driven by perceptual models and subjective criteria. The intelligent frequency processing research begins with a spectral characteristic analysis of commercial recordings, where we discover a consistent leaning towards a target equalization spectrum. A novel approach for automatically equalizing audio signals towards the observed target spectrum is then described and evaluated. We proceed to dynamics processing, and introduce an intelligent multitrack dynamic range compression algorithm, in which various audio features are proposed and validated to better describe the transient nature and spectral content of the signals. An experiment to investigate the human preference on dynamic processing is described to inform our choices of parameter automations. To provide a perceptual basis for the intelligent system, we evaluate existing perceptual models, and propose several masking metrics to quantify the masking behaviour within the multitrack mixture. Ultimately, we integrate previous research on auditory masking, frequency and dynamics processing, into one intelligent system of mix optimization that replicates the iterative process of human mixing. Within the system, we explore the relationship between equalization and dynamics processing, and propose a general frequency and dynamics processing framework. Various implementations of the intelligent system are explored and evaluated objectively and subjectively through listening experiments.

Lam Ka Man. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2004. / Includes bibliographical references (leaves 90-95). / Abstracts in English and Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Motivations --- p.3 / Chapter 1.2 --- Aims --- p.4 / Chapter 1.3 --- Contributions --- p.4 / Chapter 1.4 --- Thesis Outline --- p.5 / Chapter 2 --- Improved Recursive Modeling Method (RMM) --- p.7 / Chapter 2.1 --- An Illustrative Example --- p.7 / Chapter 2.2 --- Recursive Modeling Method (RMM) --- p.10 / Chapter 2.2.1 --- Payoff Matrices --- p.10 / Chapter 2.2.2 --- Infinite Recursive Hierarchy --- p.12 / Chapter 2.2.3 --- Choosing an Action with RMM --- p.13 / Chapter 2.3 --- Improved RMM --- p.15 / Chapter 2.3.1 --- Infinite Recursive Hierarchy --- p.15 / Chapter 2.3.2 --- The Sub-matrix Operator --- p.15 / Chapter 2.3.3 --- Finite Recursive Hierarchy --- p.17 / Chapter 2.3.4 --- Choosing an Action --- p.19 / Chapter 2.3.5 --- Recursive Formulas --- p.21 / Chapter 2.4 --- Original RMM vs. Improved RMM --- p.25 / Chapter 2.4.1 --- Terminating the Infinite Hierarchy --- p.25 / Chapter 2.4.2 --- Resultant Payoff --- p.26 / Chapter 2.5 --- Summary --- p.27 / Chapter 3 --- A Trust/Honesty Model --- p.28 / Chapter 3.1 --- The Need for a Trust Model --- p.28 / Chapter 3.1.1 --- Motivation to Tell the Truth: Invitation to Cooperate --- p.29 / Chapter 3.1.2 --- Motivation to Tell a Lie: to Prevent Competition --- p.29 / Chapter 3.1.3 --- "To Believe, or Not to Believe, that is the Question" --- p.30 / Chapter 3.2 --- The Trust Model --- p.31 / Chapter 3.2.1 --- Impression --- p.31 / Chapter 3.2.2 --- Reputation --- p.34 / Chapter 3.2.3 --- Risk Attitude and Trustworthiness --- p.34 / Chapter 3.2.4 --- Persuasiveness of a Message vs. Stubbornness of the Receiver --- p.37 / Chapter 3.3 --- The Need for an Honesty Model --- p.42 / Chapter 3.3.1 --- "To Lie, or Not to Lie, that is the Question" --- p.43 / Chapter 3.3.2 --- Problem of Living a Lie --- p.43 / Chapter 3.4 --- The Honesty Model --- p.44 / Chapter 3.4.1 --- Impression --- p.44 / Chapter 3.4.2 --- Reputation --- p.46 / Chapter 3.4.3 --- Risk Attitude and Deceivability --- p.46 / Chapter 3.4.4 --- Temptation of Lying vs. Sincerity of the Sender --- p.47 / Chapter 3.5 --- Duality of the Trust/Honesty Model --- p.52 / Chapter 3.6 --- Performance of the Trust/Honesty Model --- p.52 / Chapter 3.7 --- Summary --- p.56 / Chapter 4 --- Adaptive Strategies --- p.57 / Chapter 4.1 --- Problem of Non-adaptive Agents --- p.57 / Chapter 4.2 --- The Adaptive Strategies --- p.60 / Chapter 4.3 --- Variations of Parameters --- p.61 / Chapter 4.4 --- Adaptive Agents vs. Non-adaptive Agents --- p.63 / Chapter 4.5 --- Summary --- p.67 / Chapter 5 --- Related Work --- p.68 / Chapter 5.1 --- "Impression, Reputation and Trust" --- p.68 / Chapter 5.2 --- Theory of Honesty --- p.72 / Chapter 5.3 --- Summary --- p.72 / Chapter 6 --- Performance Analysis --- p.73 / Chapter 6.1 --- Simulation Settings --- p.73 / Chapter 6.2 --- Performance in Semi-competitive Environment --- p.75 / Chapter 6.2.1 --- Performance of Receivers --- p.75 / Chapter 6.2.2 --- Performance of Senders --- p.78 / Chapter 6.3 --- Performance when Interact with Strategic Senders --- p.80 / Chapter 6.3.1 --- Senders Telling More Truths than Lies --- p.81 / Chapter 6.3.2 --- Senders Telling More Lies than Truths --- p.83 / Chapter 6.4 --- Summary --- p.84 / Chapter 7 --- Conclusion and Future Work --- p.85 / Chapter 7.1 --- Conclusions --- p.85 / Chapter 7.2 --- Future Work --- p.87 / Chapter 7.2.1 --- Agent Manipulation --- p.87 / Chapter 7.2.2 --- Algorithm for Solving Recursive Formulas --- p.88 / Chapter 7.2.3 --- Fuzzy Trust/Honesty Model --- p.88 / Chapter 7.2.4 --- Opinion from the Mass --- p.88 / Chapter 7.2.5 --- Network Application --- p.89 / Bibliography --- p.90

Miu, Hoi Shun. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2004. / Includes bibliographical references (leaves 170-182). / Abstracts in English and Chinese. / Abstract --- p.ii / Acknowledgement --- p.v / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Problem Statement --- p.4 / Chapter 1.2 --- Contributions --- p.5 / Chapter 1.3 --- Thesis Organization --- p.6 / Chapter 2 --- Background --- p.8 / Chapter 2.1 --- Multi-agent Systems --- p.8 / Chapter 2.1.1 --- Agent Architectures --- p.9 / Chapter 2.1.2 --- Multi-agent system frameworks --- p.12 / Chapter 2.1.3 --- The Advantages and Disadvantages of Multi-agent Systems --- p.15 / Chapter 2.2 --- Evolutionary Computation --- p.16 / Chapter 2.2.1 --- Genetic Algorithms --- p.17 / Chapter 2.2.2 --- Genetic Programming --- p.18 / Chapter 2.2.3 --- Evolutionary Strategies --- p.19 / Chapter 2.2.4 --- Evolutionary Programming --- p.19 / Chapter 2.3 --- Object Recognition --- p.19 / Chapter 2.3.1 --- Knowledge Representation --- p.20 / Chapter 2.3.2 --- Object Recognition Methods --- p.21 / Chapter 2.4 --- Evolutionary Multi-agent Systems --- p.25 / Chapter 2.4.1 --- Competitive Coevolutionary Agents --- p.26 / Chapter 2.4.2 --- Cooperative Coevolutionary Agents --- p.26 / Chapter 2.4.3 --- Cellular Automata --- p.27 / Chapter 2.4.4 --- Emergent Behavior --- p.28 / Chapter 2.4.5 --- Evolutionary Agents for Image processing and Pattern Recog- nition --- p.29 / Chapter 3 --- System Architecture and Agent Behaviors in SIMULATORS --- p.33 / Chapter 3.1 --- Organization of the System --- p.34 / Chapter 3.1.1 --- General Architecture of Object Recognition System --- p.34 / Chapter 3.1.2 --- Introduction to SIMULATORS --- p.35 / Chapter 3.1.3 --- System Flow of SIMULATORS --- p.37 / Chapter 3.1.4 --- Layered Digital Image Environment --- p.39 / Chapter 3.2 --- Architecture of Autonomous Agents --- p.41 / Chapter 3.2.1 --- Internal Object Model in an Agent --- p.41 / Chapter 3.2.2 --- Current State of an Agent --- p.46 / Chapter 3.2.3 --- Local Information Sensor --- p.46 / Chapter 3.2.4 --- Direction Density Vector --- p.47 / Chapter 3.3 --- Agent Behaviors --- p.48 / Chapter 3.3.1 --- Feature Target Marking --- p.49 / Chapter 3.3.2 --- Reproduction --- p.49 / Chapter 3.3.3 --- Diffusion --- p.52 / Chapter 3.3.4 --- Vanishing --- p.54 / Chapter 3.4 --- Clustering for Autonomous Agent Training --- p.56 / Chapter 3.4.1 --- Introduction --- p.56 / Chapter 3.4.2 --- Creating the Internal Object Model --- p.58 / Chapter 3.5 --- Summary --- p.63 / Chapter 4 --- Evolutionary Algorithms for Multi Agent System --- p.64 / Chapter 4.1 --- Evolutionary Agent Behaviors in SIMULATORS --- p.65 / Chapter 4.1.1 --- Overview --- p.65 / Chapter 4.1.2 --- Evolutionary Autonomous Agents --- p.66 / Chapter 4.1.3 --- Reproduction --- p.68 / Chapter 4.1.4 --- Fitness Function --- p.68 / Chapter 4.1.5 --- Direction Density Vector Propagation --- p.73 / Chapter 4.1.6 --- Mutation --- p.73 / Chapter 4.2 --- Agents Voting Mechanism --- p.74 / Chapter 4.2.1 --- Overview --- p.74 / Chapter 4.2.2 --- Voting for Cooperative Agents --- p.75 / Chapter 4.3 --- Evolutionary Multi Agent Object Recognition --- p.79 / Chapter 4.4 --- Summary --- p.81 / Chapter 5 --- Experimental Results and Applications --- p.82 / Chapter 5.1 --- Experiment Methodology --- p.82 / Chapter 5.1.1 --- Introduction to Fung Shui Woodland --- p.83 / Chapter 5.1.2 --- Testing Images --- p.83 / Chapter 5.1.3 --- Creating Internal Object Model --- p.85 / Chapter 5.1.4 --- Experiment Parameters --- p.86 / Chapter 5.2 --- Experimental Results of Fung Shui Woodland Recognition --- p.92 / Chapter 5.2.1 --- Experiment 1: artificial0l --- p.92 / Chapter 5.2.2 --- Experiment 2: artificial0l´ؤnoise --- p.92 / Chapter 5.2.3 --- Experiment 3: artificial02 --- p.93 / Chapter 5.2.4 --- Experiment 4: FungShui0l --- p.93 / Chapter 5.2.5 --- Experiment 5: FungShui0l´ؤnoise --- p.94 / Chapter 5.2.6 --- Experiments 6 to 11: FungShui02 to FungShui07 --- p.94 / Chapter 5.3 --- Discussion --- p.119 / Chapter 5.4 --- An Example of Eyes Detection --- p.124 / Chapter 5.4.1 --- Result of the Eyes Detection --- p.128 / Chapter 5.5 --- Summary --- p.132 / Chapter 6 --- Conclusion --- p.133 / Chapter 6.1 --- Summary --- p.133 / Chapter 6.2 --- Future Work --- p.136 / Chapter A --- The Figures in the Experiments --- p.138

A Multi-agent system (MAS) is one which has a number of independent software agents interact with each other to achieve a common goal or goals automatically. It is commonly known that many species are living in the form of a MAS as it has high adaptability on surviving in an unstructural environment. Based on the benefits of group living observed from the nature, MAS is a potential direction to solve a wide range of engineering problems including robotics, chemistry, finance and genetics. / As the goodness of a response is measured from the quantity of award received, optimization is necessary in self-organization. We proposed a novel global optimization algorithm called "Creativity Driven Optimization (CDO)" in the second part of this thesis. By introducing the idea of creativity, CDO requires fewer evaluations than that of three reference methods to search for a global optimum. / In order to maximize the ability of MAS, self-organization is an essential element to be considered. Self-organization refers to a process in which the knowledge of a system accumulated automatically without being guided by an outside source or super-intelligence. In addition, the knowledge is accumulated only when the system interacts with the environment. Therefore, a robust self-organizing MAS should use a minimum number of interaction to construct the strategy for a maximum award. This goal can be achieved by involving the unstructural environment modeling and optimal response generation. / In the first part of this thesis, a radial basis function (RBF) network called "Agent Swarm Regression Network (ASRN)" is proposed in which the training algorithm is modeled as an evolution of a rule-based MAS. Three sets of experiments show that the performance of ASRN is better than that of a conventional approach in terms of computation, complexity and memory usage. The experimental results show the acceptable generalization ability and accuracy of ASRN. / In the last part of this thesis, we presented a procedure learning algorithm of self-organizing agent (PLSOA) that consists of CDO, MRN and RKL. Instead of searching for the current response with a local maximum award with reinforcement learning, PLSOA generates a response sequence by optimizing an adaptive objective function that can adjust iteratively. The experimental results of three benchmark problems show that PLSOA is able to generate nearly optimal-length response sequences in three benchmark environments. In addition, the proposed algorithm has an advantage over the reference methods in terms of reduction on procedure evaluation. / In this thesis, we have made some major contributions towards a generalized self-organizing MAS which try to mimic the MAS in nature. / To tackle the modeling process of an unstructural environment, a sequentially trained neural network called "Memory Re gression Network (MRN)" is proposed in the third part of this thesis. Based on the human's learning strategy, fewer training samples are required to train MRN in which the accuracy and generalization of MRN is similar to a reference network. After estimating the environment, the optimal response function is constructed by the estimated environment with the cooperation of a newly proposed algorithm: Response Knowledge Learning (RKL). The simulation result shows that the predator trained by RKL catches the prey within 15 steps after 50 independent successful hunting trials. / Chow Chi-kin. / "November 2005." / Advisers: H. T. Tsui; J. B. Xu. / Source: Dissertation Abstracts International, Volume: 67-11, Section: B, page: 6604. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2005. / Includes bibliographical references (p. 240-259). / 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, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstracts in English and Chinese. / School code: 1307.

Adapting to learner characteristics is essential when selecting exercises for learners in an intelligent tutoring system. This thesis investigates how humans adapt next exercise selection (in particular difficulty level) to learner personality (self-esteem), invested mental effort, and performance to inspire an adaptive exercise selection algorithm. First, we describe the investigations to produce validated materials for the main studies, namely the creation and validation of self-esteem personality stories, mental effort statements, and mathematical exercises with varying levels of difficulty. Next, through empirical studies, we investigate the impact on exercise selection of learner's selfesteem (low versus high self-esteem) and effort (minimal, little, moderate, much, and all possible effort). Three studies investigate this for learners who had different performances on a previous exercise: just passing, just failing, and performed well. Participants considered a fictional learner with a certain performance, self-esteem and effort, and selected the difficulty level of the next mathematical exercise. We found that self-esteem, mental effort, and performance all impacted the difficulty level of the exercises selected for learners. Using the results from the studies, we generated an algorithm that selects exercises with varying difficulty levels adapted to learner characteristics. Finally, through a survey with professional teachers, we evaluated our algorithm and found that the algorithm's adaptations were appropriate in general.

近年来，由于多智能体系统广泛应用于分布式传感器网络的协调和控制、移动机器人、自动驾驶车辆等，其控制设计一直是一个活跃的领域。在趋同、同步、聚类和蜂拥等问题上已经有了很多结果。线性多智能体系统的协作式输出调节问题在近几年也有所研究，但非线性多智能体系统的协作式输出调节问题的结果却很少。在这篇论文中，我们专注于非线性多智能体系统的协作式输出调节问题，分别考虑了局部情况和全局情况。 / 非线性系统的输出调节问题旨在实现轨迹跟踪和不确定非线性被控对象产生的干扰抑制，其中参考输入和干扰是通过一定的动态系统产生的，被称为外部系统。众所周知，有两种解决经典输出调节问题的方法：前馈设计和内模设计。前馈设计使用调节方程的解设计控制律，而内模设计包括两个步骤。首先把被控对象的鲁棒输出调节问题转化成由被控对象和叫做内模的动态补偿器组成的增广系统的鲁棒镇定问题，然后鲁棒镇定增广系统。 / 不同于经典的输出调节问题，协作式输出调节问题处理由N个子系统组成的多智能体系统的渐近跟踪和干扰抑制问题。N个子系统的连接关系用信息图描述。我们可以把外部系统看作领导者，把N个子系统看作外部系统的追随者。根据是否是领导者的邻居，把N个追随者分为知情的追随者和不知情的追随者。知情的追随者一组是外部系统的邻居并且可以使用自己的信息设计控制器，而不知情的追随者一组不是外部系统的邻居并且可以用其邻居的信息进行控制设计。 / 基于经典输出调节问题的这两种方法，我们利用前馈方法考虑的非线性多智能体系统的局部协作式输出调节，和通过内模方法考虑全局情况。论文的主要贡献归纳如下。 / 1. 本文考虑了非线性多智能体系统的局部协作式输出调节问题，即为，设计一个分布式控制器使得整个闭环系统当外部信号设置为零时是渐近稳定的，并且初始条件足够小时输出误差渐进趋于零。由于不知情的追随者的控制器得不到外部信号，其对应的子系统的输出调节问题不能使用自己的状态设计控制器解决。这样输出调节问题就不能用前馈设计一个分散控制器解决。因此，我们考虑协作式控制以解决输出调节问题。为了克服上述困难，我们设计了分布式控制器，包括状态反馈控制器和可测输出反馈控制器。 / 2. 本文通过协作式内膜设计研究了非线性下三角多智能体系统的全局鲁棒输出调节问题。全局鲁棒输出调节问题定义如下：找到一个控制器使得被控对象在任何初始状态下闭环系统的轨迹存在并且有界，并且对所有的初始条件，跟踪误差渐近趋近于零。有两种方法可以解决网络系统的全局鲁棒输出调节问题：分散式方法和协作式方法。通过分散式方法，对每个子系统设计一个内模，这样其控制器的阶数和子系统的数量成正比。现在，通过共享不同的追随者之间的信息，我们将利用协作式方法对所有的子系统设计只有一个内模的控制器，从而得到一个所谓的协作式控制器。这个协作式控制器的阶数和追随者的数量是独立的，并且这种方法远远比分散式方法更直接。 / 最后，我们将使用一些例子来说明我们的两种设计方法的有效性。 / In recent years, the control design of multi-agent systems has been an active area due to its wide applications in the coordination and control of distributed sensor networks, mobile robots, autonomous vehicles, etc. Many results have been obtained in such issues as consensus, synchronization, swarming and flocking. The cooperative output regulation problem for linear multi-agent systems has been studied in recent years, but there are few results on cooperative output regulation of nonlinear multi-agent systems. In this thesis, we concentrate on the cooperative output regulation problem of nonlinear multi-agent systems and consider the local case and the global case, respectively. / The output regulation problem of nonlinear systems aims to achieve asymptotic tracking and disturbance rejection in an uncertain nonlinear plant where the reference inputs and disturbances are generated by an autonomous system called the exosystem. It is known that there are two methods for solving the classical output regulation problem: feed forward design and internal model design. The feed forward design makes use of the solution of the regulator equations to design a control law while the internal model design consists of two steps. The first one is to convert the robust output regulation problem for the given plant into a robust stabilization problem for an augmented system composed of the given plant and a dynamic compensator called internal model, and the second step aims to robustly stabilize the augmented system. / Different from the classical output regulation problem, the cooperative output regulation problem handles the asymptotic tracking and disturbance rejection problem of a system consisting of N subsystems, which is the multi-agent system we consider. The connection of N subsystems is described by an information graph. We can view the exosystem as a leader system and the N subsystems as followers of the exosystem. Depending on whether or not a follower is a neighbor of the leader, the N followers can be classified into the informed followers and the uninformed followers. The group of the informed followers is the set of the neighbors of the exosystem and can use its own information for the control design, while the uninformed followers are not the neighbors of the exosystem and can use their neighbors’ information for the control design. / Based on the two approaches for studying the classical output regulation problem, we consider the local cooperative output regulation for the nonlinear multi-agent system by a feed forward approach, and the global case by an internal model approach. The main contributions are summarized as follows. / 1. The local cooperative output regulation problem for nonlinear multi-agent systems is considered, that is, design a distributed control law such that the overall closed-loop system is asymptotically stable when the exosystem signal is set to zero and the error output approaches zero asymptotically for all suciently small initial conditions. Since the control law of the uninformed followers cannot access to the exogenous signal, the output regulation problem of each uninformed follower subsystem cannot be solved by a control law utilizing its own state. Thus the output regulation problem cannot be solved by a decentralized control scheme using the feedforward design. Therefore, we consider a cooperative control to solve the out¬put regulation problem. To overcome the above diculties, the distributed control schemes are designed, including the state feedback controller and the measurement output feedback controller. / 2. The global robust output regulation problem of nonlinear lower triangular multi-agent system with uncertainties via a cooperative internal model design is studied. The global robust output regulation problem is dened as follows: nd a control law such that the trajectory of the closed-loop system starting from any initial state of the plant exists and is bounded, and the tracking error approaches zero asymptotically for all initial conditions. There are two methods to solve global robust output regulation problem of the networked systems: decentralized method and cooperative method. From decentralized method, an internal model is designed for each subsystem, which leads to a control law whose order is proportional to the number of the subsystems. Here, by sharing the information among dierent followers, we will use cooperative method and manage to design a control law with one single internal model for all subsystems, thus leads to a so-called cooperative control law. The order of the cooperative control law is independent of the number of the followers, and is much more straightforward than the decentralized method. / Finally, we will use some examples to illustrate the eectiveness of our two design methods. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Han, Qiping. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2012. / Includes bibliographical references (leaves 66-75). / Abstracts also in Chinese. / Abstract --- p.i / Acknowledgement --- p.v / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Literature Review --- p.1 / Chapter 1.2 --- Contributions of Thesis --- p.3 / Chapter 1.3 --- Thesis Outline --- p.4 / Chapter 2 --- Preliminaries --- p.6 / Chapter 2.1 --- Review of Graph Theory --- p.6 / Chapter 2.2 --- Fundamentals of Nonlinear Systems --- p.7 / Chapter 2.2.1 --- Lyapunov Stability --- p.7 / Chapter 2.2.2 --- Input-to-State Stability --- p.10 / Chapter 2.3 --- Feedforward Design of Nonlinear Output Regulation --- p.11 / Chapter 2.4 --- Internal Model Design of Nonlinear Output Regulation --- p.14 / Chapter 3 --- Local Cooperative Output Regulation --- p.22 / Chapter 3.1 --- Problem Formulation --- p.22 / Chapter 3.2 --- State Feedback Design --- p.25 / Chapter 3.3 --- Measurement Output Feedback Design --- p.34 / Chapter 3.4 --- Conclusions --- p.41 / Chapter 4 --- Global Robust Output Regulation via a Cooperative Controller --- p.42 / Chapter 4.1 --- Problem Formulation --- p.42 / Chapter 4.2 --- Solvability of the Problem --- p.49 / Chapter 4.3 --- Example --- p.54 / Chapter 4.4 --- Conclusions --- p.64 / Chapter 5 --- Conclusions --- p.65 / Bibliography --- p.66 / Biography --- p.76

Chan Wai-Chung. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2003. / Includes bibliographical references (leaves 119-121). / Abstracts in English and Chinese. / Chapter 1 --- Introduction / Chapter 1.1 --- Double Auction --- p.1 / Chapter 1.1.1 --- One-to-One Negotiation Model --- p.1 / Chapter 1.2 --- Sequential Equilibrium of One-to-One Negotiation Model --- p.3 / Chapter 1.3 --- Result --- p.4 / Chapter 2 --- Modeling the One-to-One Negotiation / Chapter 2.1 --- Nature of One-to-One Negotiation --- p.6 / Chapter 2.2 --- Basic Assumptions in the One-to-One Negotiation Model --- p.6 / Chapter 2.2.1 --- Rationality Assumption --- p.7 / Chapter 2.2.2 --- Private Valuation Assumption --- p.8 / Chapter 2.2.3 --- Subjective Belief on Opponent's Private Valuation --- p.9 / Chapter 2.3 --- Rules of the One-to-One Negotiation Model --- p.9 / Chapter 2.4 --- Payoff of Players in One-to-One Negotiation Model --- p.11 / Chapter 2.5 --- Possible Action Space of Players in One-to-One Negotiation Model --- p.12 / Chapter 2.5.1 --- Possible Action Space of the Seller Agent --- p.12 / Chapter 2.5.2 --- Possible Action Space of the Buyer Agent --- p.13 / Chapter 2.6 --- Random Vector Model for the One-to-One Negotiation Model --- p.14 / Chapter 2.6.1 --- Problems of Sequential Expectation Model --- p.14 / Chapter 2.6.2 --- Random Vector Model of the One-to-One Negotiation Game --- p.15 / Chapter 2.6.3 --- Existence of Objective Belief in Random Vector Model --- p.17 / Chapter 2.7 --- Information Set in a One-to-One Negotiation Model --- p.18 / Chapter 2.7.1 --- Game Tree of the One-to-One Negotiation Model --- p.19 / Chapter 2.7.2 --- Information Set in One-to-One Negotiation Model --- p.23 / Chapter 2.7.2.1 --- Seller's Information Set in One-to-One Negotiation Model --- p.24 / Chapter 2.7.2.2 --- Buyer's Information Set in One-to-One Negotiation Model --- p.26 / Chapter 2.8 --- Strategies of Players in One-to-One Negotiation Model --- p.28 / Chapter 2.8.1 --- Pure Strategies in the One-to-One Negotiation Model --- p.29 / Chapter 2.8.1.1 --- Payoff Function --- p.30 / Chapter 2.8.2 --- Mixed Strategies in One-to-One Negotiation Model --- p.30 / Chapter 2.8.3 --- Behavior Strategies --- p.32 / Chapter 2.9 --- Realization Probabilities in One-to-One Negotiation Model --- p.33 / Chapter 2.9.1 --- Realization Probabilities for Buyer's Information Sets and Nodes --- p.34 / Chapter 2.9.2 --- Realization Probabilities for Seller's Information Sets and Nodes --- p.35 / Chapter 2.10 --- Beliefs of Players in One-to-One Negotiation Model --- p.36 / Chapter 2.10.1 --- Seller's Belief in One-to-One Negotiation Model --- p.37 / Chapter 2.10.2 --- Buyer's Belief in One-to-One Negotiation Model --- p.38 / Chapter 2.11 --- Sequential Equilibrium of One-to-One Negotiation Model --- p.40 / Chapter 2.12 --- Applying GT for Solving Negotiation Problem --- p.41 / Chapter 3 --- Two stage One-to-One Negotiation Model / Chapter 3.1 --- Notation Used --- p.44 / Chapter 3.1.1 --- Physical Interpretation of Seller's and Buyer's Valuation --- p.44 / Chapter 3.1.2 --- Discount Factor in One-to-One Negotiation --- p.45 / Chapter 3.2 --- Formulation of Two Stage Negotiation --- p.46 / Chapter 3.2.1 --- First Stage of Negotiation Process --- p.47 / Chapter 3.2.2 --- Second Stage of Negotiation Process --- p.47 / Chapter 3.3 --- Buyer Strategies in Two Stage Negotiation --- p.49 / Chapter 3.3.1 --- Property of Equilibrium Strategy in Second Round of Negotiation --- p.49 / Chapter 3.3.2 --- Property of Equilibrium Strategy in First Round of Negotiation --- p.50 / Chapter 3.4 --- Strategic Combination of Seller Agent --- p.52 / Chapter 3.4.1 --- Three Major Types of Strategic Combination --- p.52 / Chapter 3.5 --- Properties of Type A Restricted Equilibrium Solution --- p.54 / Chapter 3.6 --- Properties of Type C Restricted Equilibrium Solution --- p.58 / Chapter 3.7 --- Properties of Type B Restricted Equilibrium Solution --- p.60 / Chapter 3.7.1 --- Relations between α1 and α2 in Type B Combinations --- p.61 / Chapter 3.7.2 --- Behavior Strategy of Buyer Agent --- p.63 / Chapter 3.7.3 --- Seller Agent's Belief in Second Round of Negotiation --- p.64 / Chapter 3.7.4 --- Seller's Payoff Function in Second Round of Negotiation --- p.65 / Chapter 3.7.5 --- Seller's Payoff Function in First Round of Negotiation --- p.67 / Chapter 3.8 --- Best Response of Seller Agent to Buyer Agent's Optimal Strategies when cb Uniformly Distributed --- p.68 / Chapter 3.8.1 --- Solutions of Type A Restricted Equilibrium Solution --- p.69 / Chapter 3.8.2 --- Solutions of Type C Restricted Equilibrium Solution --- p.71 / Chapter 3.8.3 --- Type B Restricted Equilibrium Solution of Seller Agent --- p.72 / Chapter 3.8.3.1 --- Seller's Second Round Payoff Function when cb Uniformly Distributed --- p.73 / Chapter 3.8.3.2 --- Monotonicity of Seller's Second Round Payoff Function --- p.75 / Chapter 3.8.3.3 --- Second Offer Prescribed by Equilibrium Strategy when l≥h+cs --- p.83 / Chapter 3.8.3.4 --- Second Offer Prescribed by Equilibrium Strategy when l<h+cs --- p.88 / Chapter 3.8.3.5 --- Optimization of Payoff in First Round Negotiation --- p.94 / Chapter 3.8.3.5.1 --- Type B Restricted Equilibrium Solution when l≥h+cs --- p.96 / Chapter 3.8.3.5.2 --- Type B Restricted Equilibrium Solution when l<h+cs --- p.99 / Chapter 3.9 --- Numerical Example --- p.111 / Chapter 3.9.1 --- Example 1: Type A Combination --- p.111 / Chapter 3.9.2 --- Example 2: Type B Combination --- p.113 / Chapter 3.9.3 --- Example 3: Type C Combination --- p.114 / Chapter 4 --- Conclusion and Future Works / Chapter 4.1 --- Summary of Strategies --- p.114 / Chapter 4.2 --- Future Work --- p.118 / Bibliography --- p.119