Дисертації з теми "Competing agents"

Submitted by PPG Ci?ncia da Computa??o (ppgcc@pucrs.br) on 2018-05-28T12:31:15Z No. of bitstreams: 1 TABAJARA_KRAUSBURG_RODRIGUES_DIS.pdf: 4049392 bytes, checksum: 154302eff9df959cfa74d6c0faec5d4e (MD5)
Approved for entry into archive by Sheila Dias (sheila.dias@pucrs.br) on 2018-06-06T13:05:51Z (GMT) No. of bitstreams: 1 TABAJARA_KRAUSBURG_RODRIGUES_DIS.pdf: 4049392 bytes, checksum: 154302eff9df959cfa74d6c0faec5d4e (MD5)
Made available in DSpace on 2018-06-06T13:35:29Z (GMT). No. of bitstreams: 1 TABAJARA_KRAUSBURG_RODRIGUES_DIS.pdf: 4049392 bytes, checksum: 154302eff9df959cfa74d6c0faec5d4e (MD5) Previous issue date: 2018-03-26
Esta disserta??o apresenta um estudo sobre forma??o de coaliz?es entre agentes heterog?neos para a competi??o de programa??o multiagente de 2017. Foi investigado e aplicado a forma??o de estruturas de coaliz?es entre agentes para resolver problemas log?sticos simulados sobre o mapa de uma cidade real. A fim de atingir o objetivo deste trabalho, foram integrados algoritmos formadores de coaliz?es na plataforma JaCaMo por meio de um artefato CArtAgO chamado CFArtefact. Foi utilizada a implementa??o provida pelo time SMART-JaCaMo (time participante da competi??o multiagente), para experimentar a forma??o de coaliz?es na competi??o. Tr?s abordagens foram avaliadas no dom?nio da competi??o em diferentes configura??es. A primeira abordagem utiliza somente aloca??o de tarefas para resolver o problema. A segunda e a terceira abordagem utilizam a t?cnica de forma??o de coaliz?es anteriormente ? aloca??o de tarefas; dentre estas abordagens, uma utiliza um algor?timo ?timo para resolver o problema e a outra um heur?stico. As an?lises dos experimentos realizados mostram que algor?timos formadores de coaliz?es podem melhorar a performance do time participante da competi??o quando a taxa de trabalhos gerados pelo simulador ? baixa. Entretanto, conforme a taxa de trabalhos aumenta, a abordagem que realiza somente aloca??o de tarefas obt?m um desempenho melhor quando comparada as demais. Mesmo a abordagem heur?stica tem desempenho pr?ximo ? abordagem ?tima para coaliz?es. Desta forma, ? poss?vel concluir que forma??o de coaliz?es possui grande valia para balancear os agentes para um conjunto de trabalhos que precisa ser completado.
This work focuses on coalition formation among heterogeneous agents for the 2017 multiagent programming contest. An agent is a computer system that is capable of independent action to achieve its goals. In order to increase the effectiveness of the agents, we can organise them into coalitions, in which the agents collaborate with each other to achieve individual or common goals. We investigate and apply coalition structure generation (the first activity of the coalition formation process) in simulated scenarios, specifically the 2017 contest scenario, where the agents forming a competing team cooperate to solve logistic problems simulated on the map of a real city. In order to achieve our goal, we integrate coalition formation algorithms into the JaCaMo platform by means of a CArtAgO artefact, named CFArtefact. We use the implementation of the SMART JaCaMo team for experimenting with the coalition formation approach in the contest scenario. We experiment on three approaches in the contest domain with different configurations. In the first, we use only a taskallocation mechanism, while the other approaches use an optimal coalition formation algorithm and a heuristic coalition formation algorithm. We conducted several experiments to compare the advantages of each approach. Our results show that coalition formation algorithms can improve the performance of a participating team when dealing with low job rates (i.e., how quickly new jobs are created by the simulation). However, as we increase the job rate, the approach using only task allocation has better performance. Even a heuristic coalition formation approach has close performance to the optimal one in that case. Coalition formation can play an important role when we aim to balance each group of agents to accomplish some particular goal given a larger team of cooperating agents.

Дисертація на здобуття наукового ступеня кандидата технічних наук за спеціальністю 05.13.06 – інформаційні технології. – Національний технічний університет "Харківський політехнічний інститут", Харків, 2017. Мета дисертації – підвищення ефективності процесу експертного оцінювання на основі методів, моделей та інформаційної технології підтримки процесу вибору конкуруючих агентів. Об'єкт дослідження – процес вибору конкуруючих агентів. Предмет дослідження – моделі, методи та інформаційна технологія вибору конкуруючих агентів. Розроблено нові моделі, методи та інформаційну технологію, які забез-печують підтримку впродовж всього процесу вибору конкуруючих агентів з метою підвищення якості рішення, що приймається. Запропоновано модель процесу вибору конкуруючих агентів, в якій склад-ний процес вибору декомпозується на чітко визначені етапи. Удосконалено метод визначення групових оцінок агентів на основі отримання, формалізації та агрегації індивідуальних експертних оцінок за кожним рішенням, що включає етапи абсолютного і відносного оцінювання та враховує коефіцієнти довіри до оцінки кожного учасника. Удосконалено метод оцінювання конкуруючих агентів за набором критеріїв, що дозволяє складати рейтинг агентів. Удосконалено метод оптимізації розподілу обсягу замовлення між конкуруючими агентами, який враховує відносні коефіцієнти важливості критеріїв. Створено інформаційну технологію, що дозволяє автоматизувати процес вибору конкуруючих агентів. Виконано експериментальні дослідження з вирішення практичних задач вибору конкуруючих агентів. Розроблену інформаційну систему впроваджено на підприємствах i в організаціях.
The thesis for a candidate degree in technical sciences on the specialty 05.13.06 – Information Technologies. – National Technical University "Kharkiv Polytechnic Institute", Kharkiv, 2017. The purpose of the thesis – increasing of selection process effectiveness basing on methods, models and information technology of supporting the process of competing agents selecting. The object of research – the process of competing agents selecting. The subject of research – models, methods and information technology of competing agents selecting. New models, methods and information technology which provide support of the entire process of competing agents selecting and allow to improve the quality of decisions are developed. The model of the process of competing agents selecting, which represents decomposition of the process into a set of defined stages is proposed. The method of determining group ratings based on individual expert ratings for each decision which includes obtaining of individual ratings of experts comprising absolute and relative evaluation and takes into account coefficients of confidence in the evaluations provided by each participant is improved. The method of competing agents evaluating by a set of criteria that allows to rank the agents is improved. A method of optimizing the distribution of the order among competing agents which takes into account the relative importance of criteria is improved. An information technology which allows to automate the process competing agents selecting is developed. Experimental research on solving practical problems of competing agents selection was performed. The information system is implemented in enterprises and organizations.

Дисертація на здобуття наукового ступеня кандидата технічних наук за спеціальністю 05.13.06 – інформаційні технології. – Національний технічний університет "Харківський політехнічний інститут", Харків, 2017. Мета дисертації – підвищення ефективності процесу експертного оцінювання на основі методів, моделей та інформаційної технології підтримки процесу вибору конкуруючих агентів. Об'єкт дослідження – процес вибору конкуруючих агентів. Предмет дослідження – моделі, методи та інформаційна технологія вибору конкуруючих агентів. Розроблено нові моделі, методи та інформаційну технологію, які забез-печують підтримку впродовж всього процесу вибору конкуруючих агентів з метою підвищення якості рішення, що приймається. Запропоновано модель процесу вибору конкуруючих агентів, в якій склад-ний процес вибору декомпозується на чітко визначені етапи. Удосконалено метод визначення групових оцінок агентів на основі отримання, формалізації та агрегації індивідуальних експертних оцінок за кожним рішенням, що включає етапи абсолютного і відносного оцінювання та враховує коефіцієнти довіри до оцінки кожного учасника. Удосконалено метод оцінювання конкуруючих агентів за набором критеріїв, що дозволяє складати рейтинг агентів. Удосконалено метод оптимізації розподілу обсягу замовлення між конкуруючими агентами, який враховує відносні коефіцієнти важливості критеріїв. Створено інформаційну технологію, що дозволяє автоматизувати процес вибору конкуруючих агентів. Виконано експериментальні дослідження з вирішення практичних задач вибору конкуруючих агентів. Розроблену інформаційну систему впроваджено на підприємствах i в організаціях.
The thesis for a candidate degree in technical sciences on the specialty 05.13.06 – Information Technologies. – National Technical University "Kharkiv Polytechnic Institute", Kharkiv, 2017. The purpose of the thesis – increasing of selection process effectiveness basing on methods, models and information technology of supporting the process of competing agents selecting. The object of research – the process of competing agents selecting. The subject of research – models, methods and information technology of competing agents selecting. New models, methods and information technology which provide support of the entire process of competing agents selecting and allow to improve the quality of decisions are developed. The model of the process of competing agents selecting, which represents decomposition of the process into a set of defined stages is proposed. The method of determining group ratings based on individual expert ratings for each decision which includes obtaining of individual ratings of experts comprising absolute and relative evaluation and takes into account coefficients of confidence in the evaluations provided by each participant is improved. The method of competing agents evaluating by a set of criteria that allows to rank the agents is improved. A method of optimizing the distribution of the order among competing agents which takes into account the relative importance of criteria is improved. An information technology which allows to automate the process competing agents selecting is developed. Experimental research on solving practical problems of competing agents selection was performed. The information system is implemented in enterprises and organizations.

Carrying out previously developed plans in the real world, such as picking a friend up at the airport or cooking dinner, requires that we possess some basic abilities for behaving rationally. As we all know, things just never go to plan 100 percent of the time, so possessing the facilities to cope with this uncertainty makes us good execution agents. However, the term "rational" is easy to state, and difficult to define. Just what abilities do we possess that allow us to reason in a rational manner? Researchers in the field of Artificial Intelligence (AI) have been attempting to make a precise definition of rationality since the beginning. Though we still do not have a universally accepted definition, we do have systems which to varying degrees appear to demonstrate qualities which can be said to be rational. If we characterise those systems, we can define a minimum set of capabilities which allow us to design a basic rational agent. This thesis proposes a design of a reactive execution agent (REA) which accepts and executes task directives in a dynamic environment. This design draws on an amalgam of ideas emerging from situated rational agency research. It yields an agent which is able to accept task directives (or reject them), reason about directives to determine how to achieve them (maintaining commitments to achieving other directives), respond to execution failures, and communicate with a superior agent when further deliberation is required beyond the abilities of the REA.

Human adenoviruses (Ads) as vectors have been studied for cancer gene therapy for several decades due to their ability to shut down host cell replication and lyse tumour cells. Ad5 of species C is commonly used as a replication-defective or a replication-competent vector. However, many tumour cells are relatively refractory to infection by Ad5 since the cells lack the viral receptor CAR. Thus, species B Ads are becoming more important as alternative vectors since they use CD46 as primary receptor, the expression of which is up-regulated on many tumour cell surfaces, and they also have low seroprevalence in humans. Although Ad3, Ad7, Ad11 and Ad35 have been altered to become replication-defective vectors, investigations based on replicating adenovirus vectors are still warranted.   The major aim of this thesis has been to characterize the transduction efficacy and oncolytic effect of the replication-competent adenovirus 11pGFP vector (RCAd11pGFP) in human solid tumour cell lines. Evaluation of the vector would ultimately help us to understand whether the tumour cells affect virus replication and whether the vector replicates differently in tumour cells and in untransformed diploid cells, and would eventually lead to development of more potent oncolytic adenoviruses for treatment of human cancers. The Ad11-based vector RCAd11pGFP consists of the entire Ad11p genome with a green fluorescence protein (GFP) expression cassette inserted. RCAd11pGFP shows all the characteristics of the wild-type virus and expresses GFP in cells four hours p.i. Antisera raised against Ad11p virions and hexons were able to neutralize RCAd11pGFP infection but antiserum raised against the Ad11p fibre knob could not. The infection is reduced by 90% but the fibre knob antiserum cannot completely block virus infection. Initial screening of the infection capacity of five wild-type adenoviruses in four colon cancer cell lines revealed that Ad11p, Ad11a and Ad35 of species B, showed similar replication kinetics but Ad5 showed delayed onset of virus replication in comparison to species B Ads. These data support the use of Ad11p as an alternative vector for treatment of colon cancer. The transduction efficiency of RCAd11pGFP in colon cancer and prostate cancer cell lines was studied using flow cytometry assay (FACS), and this showed that the cytolytic effect was not always in accordance with GFP expression. Toxicity assay and virus one-step replication assay showed that RCAd11pGFP replicates in highly tumorigenic cell lines (HT29, T84 and PC-3) to a greater extent than less tumorigenic cell lines (LS174T, HCT-8, DU145 and LNCaP cells), even though the latter showed relatively high GFP expression. This initial finding led to the subsequent discovery of CEACAM-family molecules, which were highly expressed in HT29 and T84 cells. Interestingly, the Ad5 wild-type virus did not manifest the same tumour-specific replication that RCAd11pGFP did in the cell lines studied. Furthermore, we investigated the influence of tumour markers for RCAd11pGFP replication in colon cancer cells. A double-staining FACS assay for detecting members of CEACAM-family molecules was established and we found that the levels of CEACAM6 were up-regulated in the cells infected by RCAd11pGFP or Ad11pwt relative to uninfected cells. However, this virus replication could not be suppressed by CEACEA6 siRNA. Our results indicate that several tumour markers or factors might be involved in promoting propagation of the virus. In vivo experiments showed significant growth inhibition of T84 and HT-29 tumours in xenograft mice treated with either RCAd11pGFP or Ad11pwt, compared to untreated controls. Furthermore, the role of the anti-tumour effect of RCAd11pGFP was also confirmed in PC3 prostate tumours in BALB/c mice. In conclusion, the novel RCAd11pGFP vector was shown to have an anti-tumour effect in vitro and in vivo. This tumour-killing effect could be enhanced in highly tumorogenic cells through virus replication. Consequently, RCAd11p may lead to development of a more potent and useful vector for human cancer therapy.

Following each of three major disasters--the financial crisis, the Gulf oil spill, and the nuclear meltdown in Japan--policymakers responded by overhauling the associated regulatory infrastructure. In each case, the response was intended to sharpen the regulator's focus, predicated on the widely held view that asking an agency to satisfy both regulatory and non-regulatory roles induces organizational conflict and impedes performance. In this dissertation, I put this commonly accepted belief about agency structure to the test by analyzing the behavior of regulators also assigned significant, non-regulatory functions. Incorporating data on a broad set of U.S. federal agencies, I first establish that the conventional wisdom holds some truth: Regulators that combine purposes do not perform as well. Even so, through a mix of statistical analyses, formal modeling, and an in-depth study of the former U.S. offshore oil and gas regulator, the Minerals Management Service, I show that assigning regulatory and non-regulatory functions to one agency can, in some cases, still be better than dividing them between agencies. I demonstrate that while the goal ambiguity and conflict introduced by combining roles does impact behavior, overemphasizing this issue misses several important factors affecting regulators tasked with non-regulatory aims. These factors explain both how regulators operate when charged with achieving other goals and why these multiple-purpose mandates persist. First, although the goals may conflict, the underlying tasks supporting these divergent purposes may still require extensive coordination. Second, even within agencies, introducing features that encourage separation between the affected groups can allow regulators to manage ambiguity, but these efforts can simultaneously exacerbate difficulties in achieving synergies generated through Professor Daniel Paul Carpenter Christopher Michael Carrigan close contact. Third, even when the conditions for conflict are present, political and public preferences—and not just internal factors—can play important roles in shaping agency priorities. Fourth, broader social, industry, and environmental shifts can attenuate or accentuate the organizational tension that exists between managing goal ambiguity while encouraging underlying coordination. In sum, only by recognizing roles for a diverse set of forces—operations, organization, politics, and environment—can the existence, behavior, and performance of regulatory agencies that balance non-regulatory mandates be logically explained.

This thesis addresses how teachers and pupils jointly constructed a physical education classroom ecology in one case study school. Taking into account the persuasive influences facing young people in contemporary society, this research addresses the complexity of young people's agendas in physical education from a socio-cultural perspective. Situating this study in a physical education class meant that I needed to explore, not only how young people's agendas influenced the classroom ecology, but also how the classroom structure influenced the agendas the young people pursued and how these were manifested during the lesson. This study is concerned primarily with describing and explaining a case of pupil engagement in physical education and to generate at a substantive level (Glaser and Strauss, 1978) theory which can make sense of the events in the one case study school. The aim is to provide some insights into why things happened the way they did and attempt to explain these happenings. The data presented in this study endorse to some extent the values claimed for the implementation of Sport Education. However, the implementation of Sport Education did not lead to all its' advocates claim for it and, although its implementation highlighted that it is possible to modify some of the agendas young people pursue during physical education lessons, I became particularly interested in the circumstances where sport education failed to hold the many benefits it purports to. Having presented these key findings, this thesis concludes with a discussion of the implications for those working with and for young people, and for the design and implementation of policies and practice in relation to physical education. Although I outline the implications of the issues emerging from this study for future practice which, at the very least, we need to consider, this study has made particularly clear the complexities of introducing change in schools.

The uplands of England are highly valued landscapes and have been powerfully shaped by agricultural practices over the centuries. Different agricultural policy discourses have impacted on the way upland farmland has been managed with both positive and negative environmental consequences. Currently, new debates are emerging about land use in the uplands, sparked by the increased interest in issues such as climate change, food security and the provision of public benefits. This thesis considered how the future of the English uplands is currently being framed at a time when these competing policy agendas are emerging. Through literature review and stakeholder interviews a number of findings have arisen. It appears that whilst food security is likely to influence upland debates to some extent, the role of climate change is likely to have a far greater influence on upland land use decisions. A key aspect that is likely to be developed is the role of the uplands in carbon and water management. This thesis drew upon different scenarios for the future of the upland and asked which are the most likely or desirable in the light of changing policy interests. There appear to be a number of possible futures for the uplands depending on how relevant policies evolve. The most desired future among stakeholders is one in which the uplands become viewed as ‘Environmentally Favoured Areas’, with an emphasis on the ecosystem services they provide. A clear need is highlighted for further research into how to value and provide for such services.

The purpose and aim of this study is to offer a hitherto lacking perspective on the Sulu Archipelago’s history during the early years of American colonial rule – that of the native Sulu Sultan. Existing research, albeit new, tends to favour a U.S centred view. With concurrences as a theoretical framework and a theoretical understanding of the social and political power dynamics in the Sultanate, this study aims to challenge that perspective by examining the last Sultan of Sulu, Jamalul Kiram II’s agency and room for manoeuvre in relation to the bilateral Kiram-Bates treaty between the Sultanate of Sulu in the southern Philippines and the United States, represented by Colonel John C. Bates. Through a qualitative analysis of letters, official documents and a memoire originating from, or regarding, the time period, this study indicates that the Sultan’s agency and room for manoeuvre in relation to the Kiram-Bates treaty has been overlooked by scholars, and his position somewhat poorly understood. The Sultan did not simply sign the treaty according to American wishes. He negotiated terms, and influenced the outcome of above all the signing, and he at least reacted to and navigated the abrogation of the agreement in a way that has not been sufficiently described before. These findings prompt further research from a concurrences perspective to more thoroughly understand the history, and perhaps even the contemporary state, of the Sulu Archipelago and the Sulu Sultanate.

Building upon feminist and sexual health research, this dissertation shows how the positioning of women in various discourses as somehow ‘lacking’ actually constrains what researchers are able to hear in their sexual stories. Using interviews with 26 heterosexually active young women, I seek to upset traditional approaches to understanding young women’s sexual stories and theorizing heterosexuality. To analyze the interviews, I first employ a Foucauldian-inspired discourse analysis that focuses on the power that circulates through discourses and our positioning within them. Our positioning in various discourses both enables and limits various courses of action, understandings and experiences. This power of discourse is illustrated by an emergent hybrid discourse that is apparent in young women’s sexual narratives. I discuss what I call the ‘competent feminine sexuality’ discourse and show how this discourse smoothes over contradictions between liberal and gendered discourses. Secondly, I show how psychoanalytic insights allow us to explore the processes of subjectification by which young women constitute themselves as (hetero)sexual women. Specifically, this dissertation explores processes of abjection, disavowal and ambivalence in participants’ narratives. In conclusion, the dissertation outlines the practical implications for sexual health education in Canada.

The growing interest in sustainability amounts pressure on organisations to operate in more environmentally friendly ways, sparking the need for radical sustainable change. The purpose of this study is to discover what factors and conditions facilitate and hinder organisational learning for sustainability transitions (ST), a topic that has caught recent academic attention and thus needs further interdisciplinary research. The conceptual framework derives inspiration from the Four Factors of Learning for ST whilst the Competing Values Framework and the Three Loops of Learning are used as additional lens to better understand the conditions of learning for ST. The thesis will take a qualitative approach through conducting a content analysis of three company documents and seven semi-structured interviews with employees from a public service agency in Scandinavia currently implementing a transition lab. A descriptive analysis of the coded data highlights the levels of understanding and acceptance towards sustainability transitions from the case organisation and the conditions that facilitate and hinder organisational learning. The results show that the most prominent of the Four Factors of Learning for ST in the early phase of a sustainability transition is interpersonal, followed by material, institutional and intrapersonal. Within these factors, the discussion further highlights the most prevalent sub-codes and themes that reoccur in the data. Moreover, five key findings under the themes of flexibility versus control, resource availability through digitalisation, communication, collaboration and facing complexity using institutional logics were identified as the primary facilitating and hindering factors that promote learning for STs. Finally, recommendations are presented to inform both theory, and practice, as further analysing learning for ST is of high relevance to better understand and design these learning journeys and a more sustainable (organisational) future.

碩士
逢甲大學
統計與精算所
100
Scheduling with two competing agents has become a popular research topic in recent years. Most of the research focuses on the objective functions with constraint forms, that is, to minimize the objective function of jobs from one agent given that the objective function of jobs from the other agent does not exceed an upper bound. In this paper, we consider a different approach for a two-agent problem on a single machine. The objective is the weighted sum of the total completion time of jobs from one agent and the total tardiness of jobs from the other agent. Two branch-and-bound algorithms are proposed to derive the optimal sequences. In addition, a simulated annealing and two genetic algorithms are provided to search for the near-optimal solutions. Computational experiments show that the proposed branch-and branch algorithm could solve problems of up to 40 jobs in a reasonable amount of time and the genetic algorithm that use initial sequence is good with an average error percentage of less then 1.5% for all the tested cases.

Urban greening has evolved into a diverse set of strategies based on community organizing and policy change to help create healthy communities. Currently, there are competing urban greening agendas and barriers to coalition building that prohibit urban greening projects and policies from reaching their maximum potential. Greening projects take the collective effort of residents, city government, nonprofits, community-based organizations and a range of technical experts. To move the agenda forward there is a need to create a framework around health and build community-based health coalitions. A healthy community framework taps into universal concerns and the need to build sustainable communities. In order to move forward we need collective action, coalition building and grassroots organizing in conjunction with economics, science, policy, planning, design and law.
text

Yip Kin Fung.
Thesis (M.Phil.)--Chinese University of Hong Kong, 2003.
Includes bibliographical references (leaves 101-104).
Abstracts in English and Chinese.
Chapter 1 --- Introduction --- p.1
Chapter 2 --- The Distribution of Fluctuations in Financial Data --- p.5
Chapter 2.1 --- Empirical Statistics --- p.5
Chapter 2.2 --- Data analyzed --- p.8
Chapter 2.3 --- Levy Distribution --- p.10
Chapter 2.4 --- Returns Distribution and Scaling Properties --- p.12
Chapter 2.5 --- Volatility Clustering --- p.19
Chapter 2.6 --- Conclusion --- p.21
Chapter 3 --- Models of Herd behaviour in Financial Markets --- p.22
Chapter 3.1 --- Cont and Bouchaud's model --- p.22
Chapter 3.2 --- The Model of Egiuluz and Zimmerman --- p.24
Chapter 3.3 --- EZ Model with Size-Dependent Dissociation Rates --- p.28
Chapter 3.4 --- Democratic and Dictatorship Self-Organized Model --- p.31
Chapter 3.5 --- Effect of Size-Dependent Fragmentation and Coagulation Prob- abilities --- p.33
Chapter 3.6 --- Extensions of EZ model --- p.35
Chapter 3.7 --- Conclusion --- p.39
Chapter 4 --- Review on the Minority Game(MG) --- p.42
Chapter 4.1 --- The Model and Results --- p.42
Chapter 4.2 --- Crowd-anticrowd Theory and Phase Transition --- p.46
Chapter 4.3 --- Market Efficiency --- p.48
Chapter 5 --- MG with biased strategy pool --- p.52
Chapter 5.1 --- The Model --- p.53
Chapter 5.2 --- Numerical Results and Discussion --- p.53
Chapter 5.3 --- Theory: MG with Biased Strategy Pool --- p.61
Chapter 5.4 --- Conclusion --- p.69
Chapter 6 --- MG with Randomly Participating Agents --- p.71
Chapter 6.1 --- The Model with One RPA --- p.71
Chapter 6.2 --- Results for q = 0.5 --- p.72
Chapter 6.3 --- Inefficiency and Success Rate --- p.76
Chapter 6.4 --- Results for q ≠ 0.5 --- p.82
Chapter 6.5 --- Many RPAs --- p.85
Chapter 6.6 --- Conclusion --- p.86
Chapter 7 --- A Model on Coupled Minority Games --- p.88
Chapter 7.1 --- The Model --- p.89
Chapter 7.2 --- Results and Discussion。 --- p.90
Chapter 7.3 --- Conclusion --- p.95
Chapter 8 --- Conclusion --- p.97
Bibliography --- p.101
Chapter A --- Solving Cluster Size distribution in EZ model --- p.105

Chan Hoi-Yeung = 網絡與競爭系統的物理 : 個體為本模型中的網絡效應 / 陳凱揚.
Thesis submitted in: September 2005.
Thesis (M.Phil.)--Chinese University of Hong Kong, 2006.
Includes bibliographical references (leaves 191-197).
Text in English; abstracts in English and Chinese.
Chan Hoi-Yeung = Wang luo yu jing zheng xi tong de wu li : ge ti wei ben mo xing zhong de wang luo xiao ying / Chen Kaiyang.
Abstract --- p.i
Acknowledgments --- p.v
Contents --- p.vii
Chapter 1 --- Overview --- p.1
Chapter I --- Networks --- p.3
Chapter 2 --- Networks in nature --- p.4
Chapter 2.1 --- Introduction --- p.4
Chapter 2.2 --- Terminology of the networks studies --- p.6
Chapter 2.2.1 --- Nodes --- p.6
Chapter 2.2.2 --- Links --- p.6
Chapter 2.2.3 --- Adjacency matrix --- p.9
Chapter 2.2.4 --- Connectivity --- p.10
Chapter 2.2.5 --- Clustering coefficient --- p.11
Chapter 2.2.6 --- Shortest path --- p.11
Chapter 2.2.7 --- Connectivity correlation --- p.12
Chapter 2.3 --- Topology in the real-world networks --- p.13
Chapter 2.3.1 --- The Internet --- p.13
Chapter 2.3.2 --- The WWW --- p.15
Chapter 2.3.3 --- Collaboration networks --- p.15
Chapter 2.3.4 --- Food webs --- p.16
Chapter 2.3.5 --- Power grids --- p.17
Chapter 2.4 --- Discussion --- p.17
Chapter 3 --- Review on Network Models --- p.19
Chapter 3.1 --- Introduction --- p.19
Chapter 3.2 --- Graph Theory --- p.20
Chapter 3.2.1 --- Classical random graph --- p.20
Chapter 3.3 --- Evolving networks --- p.23
Chapter 3.3.1 --- Random growing network --- p.23
Chapter 3.3.2 --- Fitness growing network --- p.25
Chapter 3.3.3 --- Barabasi-Albert model --- p.27
Chapter 3.3.4 --- Fitness model --- p.31
Chapter 3.4 --- Lattice --- p.33
Chapter 3.4.1 --- Regular hypercubic lattices (Periodic) --- p.33
Chapter 3.4.2 --- Regular hypercubic lattices (Free boundary conditions) . --- p.35
Chapter 3.5 --- Discussion --- p.35
Chapter 4 --- Network Properties --- p.38
Chapter 4.1 --- More derivations on existing models --- p.38
Chapter 4.1.1 --- Classical random graphs --- p.38
Chapter 4.1.2 --- Barabasi-Albert model --- p.40
Chapter 4.1.3 --- Fitness Model --- p.42
Chapter 4.1.4 --- Regular hypercubic lattices (Periodic) --- p.45
Chapter 4.2 --- New model --- p.48
Chapter 4.2.1 --- Fitness-BA hybrid model --- p.48
Chapter 4.3 --- Link removal --- p.55
Chapter 4.3.1 --- Introduction --- p.55
Chapter 4.3.2 --- Formalism in connectivity --- p.55
Chapter 4.3.3 --- Pruned BA Model --- p.56
Chapter 4.4 --- Link addition --- p.58
Chapter 4.4.1 --- Introduction --- p.58
Chapter 4.4.2 --- Regular hypercubic lattices (Periodic) --- p.58
Chapter 4.5 --- Discussion --- p.60
Chapter II --- Games --- p.62
Chapter 5 --- Review on Agent-based models of competing population --- p.63
Chapter 5.1 --- Introduction --- p.63
Chapter 5.2 --- The El Farol Bar attendance problem --- p.65
Chapter 5.2.1 --- Model --- p.65
Chapter 5.2.2 --- Strategies --- p.66
Chapter 5.2.3 --- Discussion --- p.66
Chapter 5.3 --- Minority game --- p.67
Chapter 5.3.1 --- Model --- p.67
Chapter 5.3.2 --- Strategies --- p.68
Chapter 5.3.3 --- Attendance --- p.69
Chapter 5.3.4 --- History and quasi-Eulerian state --- p.69
Chapter 5.3.5 --- Success rate and Hamming distance --- p.71
Chapter 5.3.6 --- Volatility --- p.73
Chapter 5.3.7 --- Crowd-anticrowd theory --- p.75
Chapter 5.3.8 --- Discussion --- p.76
Chapter 6 --- B-A-R model : Dynamics --- p.78
Chapter 6.1 --- Model --- p.78
Chapter 6.2 --- Results: Plateaux and periodicity --- p.81
Chapter 6.3 --- A microscopic view: Agents' decisions and strategy performance --- p.86
Chapter 6.4 --- A macroscopic view: Bit-string patterns --- p.92
Chapter 6.4.1 --- The history space --- p.92
Chapter 6.4.2 --- Bit-string statistics of different states --- p.94
Chapter 6.5 --- The (max = 1 states --- p.97
Chapter 6.5.1 --- Values of wm3iX --- p.97
Chapter 6.5.2 --- "Strategy ranking evolvement: ni, (w)" --- p.101
Chapter 6.5.3 --- Substates . --- p.105
Chapter 7 --- B-A-R model : Formalism --- p.108
Chapter 7.1 --- Resource level at transitions of Cmax = 0 state --- p.108
Chapter 7.2 --- Resource levels at transitions of Cmax 二 1 states --- p.109
Chapter 7.2.1 --- Method --- p.109
Chapter 7.2.2 --- Lmin for upper substate --- p.110
Chapter 7.2.3 --- Lmin for lower substate --- p.113
Chapter 7.3 --- Discussion --- p.116
Chapter 8 --- B-A-R model : Statistics --- p.121
Chapter 8.1 --- Problem --- p.121
Chapter 8.2 --- Bit-string statistics --- p.122
Chapter 8.2.1 --- Allowed transitions --- p.122
Chapter 8.2.2 --- Grouping the history space --- p.122
Chapter 8.2.3 --- "Grouping the states, Cmax" --- p.127
Chapter 8.2.4 --- "Labelling each state, /(C)" --- p.129
Chapter 8.3 --- Discussion --- p.130
Chapter III --- Networked games --- p.131
Chapter 9 --- Networked minority game --- p.132
Chapter 9.1 --- Model --- p.132
Chapter 9.2 --- Preliminary results: Agents' success rates --- p.133
Chapter 9.3 --- Ranking the strategies --- p.135
Chapter 9.3.1 --- Ranking pattern --- p.136
Chapter 9.3.2 --- Fraction of strategies in each rank --- p.140
Chapter 9.4 --- Number of agents using a best strategy belonging to rank r --- p.141
Chapter 9.4.1 --- Unconnected population --- p.141
Chapter 9.4.2 --- Networked population . --- p.142
Chapter 9.5 --- Application: Mean success rate --- p.143
Chapter 9.6 --- Mean success rate of agents with degree k --- p.147
Chapter 9.7 --- Application in other networks --- p.149
Chapter 9.8 --- Discussion --- p.151
Chapter 10 --- Interacting agents: Networked B-A-R model --- p.154
Chapter 10.1 --- Model --- p.154
Chapter 10.2 --- The quasi-Eulerian state (wmax = 1/2 state) --- p.155
Chapter 10.3 --- The emergent states --- p.159
Chapter 10.3.1 --- General results --- p.159
Chapter 10.3.2 --- The Cmax = 0 state --- p.160
Chapter 10.3.3 --- The Cmax = 1 state --- p.161
Chapter 10.4 --- Discussion --- p.162
Chapter IV --- Conclusion --- p.164
Chapter 11 --- Conclusion --- p.165
Chapter V --- Appendices --- p.172
Chapter A --- List of symbols --- p.173
Chapter A.1 --- Networks --- p.173
Chapter A.2 --- Games --- p.174
Chapter A.3 --- Networked games --- p.176
Chapter B --- Distance distribution in classical random graphs --- p.177
Chapter B.1 --- Method --- p.177
Chapter B.2 --- Distance distribution --- p.177
Chapter B.3 --- Behaviour at small L --- p.178
Chapter B.4 --- Behaviour at large L --- p.179
Chapter C --- Co-ordination number in infinite hypercubic lattice --- p.181
Chapter C.1 --- Method --- p.181
Chapter C.1.1 --- ID lattice --- p.181
Chapter C.1.2 --- 2D square lattice --- p.182
Chapter C.1.3 --- Higher dimension hypercubic lattices --- p.183
Chapter C.2 --- Coefficients --- p.185
Chapter D --- Connectivity distribution in fitness-BA hybrid model --- p.187
Chapter D.1 --- Mean field approach --- p.187
Chapter D.2 --- Connectivity distribution --- p.188
Chapter D.3 --- Power-law exponent --- p.190
Bibliography --- p.191

Chan King Pak Keven = 競爭性系統個體模型中的策略動態、決策及整體表現 / 陳景柏.
Thesis submitted in: August 2005.
Thesis (M.Phil.)--Chinese University of Hong Kong, 2006.
Includes bibliographical references (leaves vii-viii (4th gp.)).
Text in English; abstracts in English and Chinese.
Chan King Pak Keven = Jing zheng xing xi tong ge ti mo xing zhong de ce lüe dong tai, jue ce ji zheng ti biao xian / Chen Jingbo.
Chapter 1 --- Introduction --- p.1
Chapter 2 --- Review on the Minority Game --- p.5
Chapter 2.1 --- Background --- p.5
Chapter 2.2 --- Model of MG --- p.6
Chapter 2.3 --- Features --- p.7
Chapter 2.3.1 --- Phase Transition --- p.7
Chapter 2.3.2 --- Inefficient and Efficient Phase --- p.8
Chapter 2.3.3 --- Anti-persistence --- p.9
Chapter 2.3.4 --- Data Collapse --- p.10
Chapter 2.4 --- Existing Theories --- p.10
Chapter 2.4.1 --- Reduced Strategy Space --- p.11
Chapter 2.4.2 --- The Crowd-Anticrowd Theory --- p.12
Chapter 2.5 --- Summary --- p.13
Chapter 3 --- Introduction to Strategy Ranking Theory --- p.15
Chapter 3.1 --- Strategy Ranking Theory for Mean Success Rate --- p.15
Chapter 3.1.1 --- Time evolution of Virtual Point Ranking --- p.15
Chapter 3.1.2 --- Winning Probability for m = 1 --- p.17
Chapter 3.2 --- Calculation of Mean Success Rate --- p.21
Chapter 3.3 --- "Size Dependence of weυen(K) (""Market Impact"" Effect)" --- p.23
Chapter 3.4 --- Size Dependence of wodd、K) (Uneven Distribution of Agents into Split Ranks) --- p.25
Chapter 4 --- Implementation of Strategy Ranking Theory --- p.30
Chapter 4.1 --- Feature of wodd(k) for higher m --- p.30
Chapter 4.2 --- Derivation of wodd(k) from Strategy Ranking Theory --- p.32
Chapter 4.3 --- Proof of Eq. (4.14) --- p.36
Chapter 4.4 --- Discussion on wodd(k) --- p.41
Chapter 4.4.1 --- Asymptotic Behavior of wodd(k) --- p.42
Chapter 4.4.2 --- Finite size correction of wodd(k) --- p.43
Chapter 5 --- Applications of Strategy Ranking Theory --- p.46
Chapter 5.1 --- Probability Density Function of Agents Making a Particular Choice --- p.46
Chapter 5.1.1 --- Odd time steps： k = 1 --- p.47
Chapter 5.1.2 --- Odd time steps： k = 2 --- p.48
Chapter 5.1.3 --- "Rodd,K" --- p.49
Chapter 5.1.4 --- Even time steps --- p.51
Chapter 5.1.5 --- Overall Attendance Distribution --- p.51
Chapter 5.2 --- The Variance of the Attendance --- p.52
Chapter 5.2.1 --- Asymptotic behavior of the variance --- p.54
Chapter 5.3 --- Anti-persistent Nature of Efficient Phase of MG --- p.55
Chapter 5.4 --- Summary --- p.58
Chapter 6 --- Strategy Ranking Theory and Crowd-Anticrowd Theory --- p.59
Chapter 6.1 --- Introduction --- p.59
Chapter 6.1.1 --- Strategy Ranking Theory --- p.60
Chapter 6.1.2 --- Crowd-Anticrowd Theory --- p.61
Chapter 6.2 --- Crowd-Anticrowd Theory with Ranking Patterns Characterized by k --- p.63
Chapter 6.3 --- Variance: Crowd-Anticrowd Theory --- p.65
Chapter 6.3.1 --- m = 1 --- p.65
Chapter 6.3.2 --- m = 2 --- p.66
Chapter 6.4 --- Variance： Modified Crowd-Anticrowd Theory for m̐ơح 1 --- p.66
Chapter 6.4.1 --- k = 0 --- p.67
Chapter 6.4.2 --- k = 1 --- p.67
Chapter 6.4.3 --- k = 2 --- p.67
Chapter 6.4.4 --- Sum over all k --- p.68
Chapter 6.5 --- Variance: Modified Crowd-Ant icrowd Theory for m=2 --- p.68
Chapter 6.5.1 --- k = 3 --- p.69
Chapter 6.5.2 --- k = 4 --- p.70
Chapter 6.5.3 --- Sum over all k --- p.71
Chapter 6.6 --- "Strategy Ranking Theory Expressed in (Nkl-Nk,(l)" --- p.71
Chapter 6.7 --- Summary --- p.73
Chapter 7 --- Variance of the Attendance in MG: Data Collapse --- p.75
Chapter 7.1 --- Previous Studies --- p.75
Chapter 7.2 --- Attempt 1 --- p.76
Chapter 7.2.1 --- Understanding from the Existing Theories --- p.76
Chapter 7.2.2 --- Numerical Results --- p.79
Chapter 7.3 --- Attempt 2 --- p.80
Chapter 7.3.1 --- Modification Based on αc ß 1/2 --- p.81
Chapter 7.3.2 --- Numerical Results --- p.81
Chapter 7.4 --- Summary --- p.82
Chapter 8 --- Minority Game in Networked Population --- p.83
Chapter 8.1 --- Introduction --- p.83
Chapter 8.2 --- Model --- p.84
Chapter 8.3 --- Numerical Results --- p.85
Chapter 8.4 --- Classification of Predictors --- p.86
Chapter 8.4.1 --- Major Classification of Predictors - Hamming Distance D --- p.87
Chapter 8.4.2 --- "Minor Classification of Predictors - Dynamical Ranking (k,1)" --- p.88
Chapter 8.4.3 --- "Using the Classification (k,l, D)" --- p.89
Chapter 8.5 --- "Winning Probability of a Predictor (wk,l,d)" --- p.89
Chapter 8.5.1 --- "Odd Steps, k = 1" --- p.90
Chapter 8.5.2 --- "Odd Steps, k = 2" --- p.91
Chapter 8.6 --- Number of Predictors --- p.93
Chapter 8.7 --- Mean Success Rate of Non-networked MG： m = 1 --- p.93
Chapter 8.8 --- "Cluster Size of a Predictor (sk,l,D)" --- p.95
Chapter 8.9 --- Mean Success Rate of Networked MG --- p.97
Chapter 8.9.1 --- With wK(even)=0.5 --- p.97
Chapter 8.9.2 --- "Modification of wK(even) Using skl,D" --- p.98
Chapter 8.9.3 --- Modification of Using Modified wK(even) --- p.100
Chapter 8.10 --- Variance of the Attendance in Networked MG --- p.101
Chapter 8.11 --- Attendance Distribution --- p.103
Chapter 8.12 --- A Network-type Independent Approach --- p.104
Chapter 8.12.1 --- Degree Depending Success Rate --- p.104
Chapter 8.12.2 --- Evaluating w(k) --- p.107
Chapter 8.12.3 --- Application on Random Graph as Underlying Network --- p.108
Chapter 8.13 --- The Position of the Minimum Variance --- p.108
Chapter 8.14 --- Summary --- p.110
Chapter 9 --- Conclusion --- p.111
Bibliography --- p.115

碩士
中央警察大學
水上警察研究所
107
According to Fisheries Act, the Fishery Competent Authority is responsible for the maritime safety of fishing vessels. In 2000, Coast Guard Agency is established and enforces port security inspections in accordance with Coast Guard Act. First On November 28th 2012, Article 11-1 of Fisheries Act has been amended and published where any fishing vessel leaving port in violation of the provisions, the Fishery Competent Authority may commission the Coast Guard Agency to take appropriate measures to prevent the vessel from leaving port, and may take compulsory measures, it’s “commission-enforce”. And on July 20th 2016, Article 49 of Fisheries Act has been amended and published that the Coast Guard Agency may pursuant to its mandate, designate officer(s) to the fishing vessel of the fishery operator for inspection and question any relevant party. The relevant party shall not evade, obstruct or deny any of the inspections, it’s “enforce by law”. The difference of them is “Coast Guard Administration, Executive Yuan” be established, it not only confirmed the legal enforcement system about combination of coast and sea, but also protected the Nation’s right of the sea and people’s life property, focus on the appropriateness of law enforcement in two aspects. In the process of law enforcement, shall take into account the principles of fairness, appropriateness, and thoroughness, actively develop towards the ocean and create a new era of maritime and coastal patrols in Taiwan. The Legislative Yuan’s reason for the revision of the law in 2016, updated “The Coast Guard Agency may designate officer(s) to the fishing vessel for inspection and question to meet practical needs.” To see the importance of the Coast Guard Agency for fisheries law enforcement. However, Fisheries Act’s central competent authority is Council of Agriculture, Executive Yuan; in fact, the purpose of the business is the organization of its three levels of institutions: Fishery Agency that has jurisdiction with Coast Guard Agency for each one. There’re two or more ways to enforce in interrelated systems, it’s controversial to enforce law of fishing vessels. This study reference related legal system of Japan especially, trying to explore the duties and divisions between the Fisheries Agency and the Coast Guard Agency, and review whether the enforcement methods of the Coast Guard Agency are safe in practice, suggest amendments and what the Coast Guard Agency should do.