This thesis is concerned with policy space abstractions that concisely encode alternative ways of making decisions; dealing with discovery, learning, adaptation and use of these abstractions. This work is motivated by the problem faced by autonomous agents that operate within a domain for long periods of time, hence having to learn to solve many different task instances that share some structural attributes. An example of such a domain is an autonomous robot in a dynamic domestic environment. Such environments raise the need for transfer of knowledge, so as to eliminate the need for long learning trials after deployment. Typically, these tasks would be modelled as sequential decision making problems, including path optimisation for navigation tasks, or Markov Decision Process models for more general tasks. Learning within such models often takes the form of online learning or reinforcement learning. However, handling issues such as knowledge transfer and multiple task instances requires notions of structure and hierarchy, and that raises several questions that form the topic of this thesis – (a) can an agent acquire such hierarchies in policies in an online, incremental manner, (b) can we devise mathematically rigorous ways to abstract policies based on qualitative attributes, (c) when it is inconvenient to employ prolonged trial and error learning, can we devise alternate algorithmic methods for decision making in a lifelong setting? The first contribution of this thesis is an algorithmic method for incrementally acquiring hierarchical policies. Working with the framework of options - temporally extended actions - in reinforcement learning, we present a method for discovering persistent subtasks that define useful options for a particular domain. Our algorithm builds on a probabilistic mixture model in state space to define a generalised and persistent form of ‘bottlenecks’, and suggests suitable policy fragments to make options. In order to continuously update this hierarchy, we devise an incremental process which runs in the background and takes care of proposing and forgetting options. We evaluate this framework in simulated worlds, including the RoboCup 2D simulation league domain. The second contribution of this thesis is in defining abstractions in terms of equivalence classes of trajectories. Utilising recently developed techniques from computational topology, in particular the concept of persistent homology, we show that a library of feasible trajectories could be retracted to representative paths that may be sufficient for reasoning about plans at the abstract level. We present a complete framework, starting from a novel construction of a simplicial complex that describes higher-order connectivity properties of a spatial domain, to methods for computing the homology of this complex at varying resolutions. The resulting abstractions are motion primitives that may be used as topological options, contributing a novel criterion for option discovery. This is validated by experiments in simulated 2D robot navigation, and in manipulation using a physical robot platform. Finally, we develop techniques for solving a family of related, but different, problem instances through policy reuse of a finite policy library acquired over the agent’s lifetime. This represents an alternative approach when traditional methods such as hierarchical reinforcement learning are not computationally feasible. We abstract the policy space using a non-parametric model of performance of policies in multiple task instances, so that decision making is posed as a Bayesian choice regarding what to reuse. This is one approach to transfer learning that is motivated by the needs of practical long-lived systems. We show the merits of such Bayesian policy reuse in simulated real-time interactive systems, including online personalisation and surveillance.

A presente tese se propõe a responder a seguinte questão: É correto afirmar que o regionalismo desenvolvimentista sul-americano, implantado ao longo do século XX, e o novo regionalismo desenvolvimentista, implantado ao longo da primeira década do século XXI, são compatíveis com as regras do sistema multilateral de comércio? Em outras palavras, as organizações de integração regional da América do Sul têm balizado as suas políticas desenvolvimentistas no policy space decorrente das prerrogativas jurídicas para o desenvolvimento do sistema multilateral de comércio? Para tratar o refrido tema, parte-se da premissa que os países da América do Sul historicamente sempre fizeram uso das prerrogativas jurídicas para o desenvolvimento, desde que tais surgiram no sistema multilateral de comércio, e possibilitaram a implantação de políticas no plano nacional e regional. Ao longo da tese, é analisada, desde uma perspectiva jurídica, a compatibilidade com as regras do sistema multilateral de comércio dos principais aspectos desenvolvimentistas das organizações de integração regional criadas na América do Sul, desde a ALALC, em 1960, até a UNASUL, em 2008. Concluiu-se que tanto o velho quanto o novo regionalismo desenvolvimentista da América do Sul foram levados adiante em concordância com o direito internacional do comércio, seja ao longo do período GATT, seja ao longo do período OMC. Mesmo com a redução de policy space decorrente da entrada em vigor dos acordos da OMC, a América do Sul vislumbrou manter um relativamente alto nível de espaço político (policy space) para a criação de políticas públicas de desenvolvimento. Atualmente, percebe-se uma tendência das organizações de integração regional da América do Sul, especialmente da UNASUL e do MERCOSUL, de levar adiante políticas públicas em torno de temas não-regulados pela OMC, o que caracterizaria uma estratégia OMC-extra. O novo regionalismo desenvolvimentista sul-americano está finalmente se constituindo para além do policy space do sistema multilateral de comércio, o que, por não configurar, a priori, uma incompatibilidade dos processos de integração com as regras da OMC, contribui para fortalecer o primado do direito internacional na América do Sul. / This thesis aims to answer the following question: Can it be said that the South American developmental regionalism, carried out throughout the twentieth century, and the new developmental regionalism, implemented during the first decade of this century, are compatible with the rules of multilateral trading system? In other words, have the regional integration organizations in South America based their developmental policies in the policy space resulting from the developmental legal prerogatives of the multilateral trading system? To treat such theme, we start from the premise that the countries of South America have historically made use of legal rights for development, since these arose in the multilateral trading system, and enabled the implementation of policies at the national and regional levels. From a legal perspective, the thesis analyses the compatibility of the main developmental aspects of regional integration organizations created in South America since the LAFTA in 1960 to UNASUR, in 2008 with the rules of the multilateral trading system. It was concluded that both the old and the new developmental regionalism in South America were brought forward in accordance with international trade law, either through the GATT period, either during the WTO period. Even with the reduction of policy space resulting from the entry into force of the WTO agreements, South America managed to maintain a relatively high level of policy space for the creation of developmental public policies. Currently, there is a perceived trend of regional integration organizations in South America, especially the MERCOSUR and UNASUR, to carry out public policies on topics unregulated by the WTO, which would characterize a strategy WTO-extra. The new South American developmental regionalism is finally going beyond the policy space of the multilateral trading system, which, by not setting a priori incompatibility of the integration processes with WTO rules, contributes to strengthen the rule of international law in South America.

A presente tese se propõe a responder a seguinte questão: É correto afirmar que o regionalismo desenvolvimentista sul-americano, implantado ao longo do século XX, e o novo regionalismo desenvolvimentista, implantado ao longo da primeira década do século XXI, são compatíveis com as regras do sistema multilateral de comércio? Em outras palavras, as organizações de integração regional da América do Sul têm balizado as suas políticas desenvolvimentistas no policy space decorrente das prerrogativas jurídicas para o desenvolvimento do sistema multilateral de comércio? Para tratar o refrido tema, parte-se da premissa que os países da América do Sul historicamente sempre fizeram uso das prerrogativas jurídicas para o desenvolvimento, desde que tais surgiram no sistema multilateral de comércio, e possibilitaram a implantação de políticas no plano nacional e regional. Ao longo da tese, é analisada, desde uma perspectiva jurídica, a compatibilidade com as regras do sistema multilateral de comércio dos principais aspectos desenvolvimentistas das organizações de integração regional criadas na América do Sul, desde a ALALC, em 1960, até a UNASUL, em 2008. Concluiu-se que tanto o velho quanto o novo regionalismo desenvolvimentista da América do Sul foram levados adiante em concordância com o direito internacional do comércio, seja ao longo do período GATT, seja ao longo do período OMC. Mesmo com a redução de policy space decorrente da entrada em vigor dos acordos da OMC, a América do Sul vislumbrou manter um relativamente alto nível de espaço político (policy space) para a criação de políticas públicas de desenvolvimento. Atualmente, percebe-se uma tendência das organizações de integração regional da América do Sul, especialmente da UNASUL e do MERCOSUL, de levar adiante políticas públicas em torno de temas não-regulados pela OMC, o que caracterizaria uma estratégia OMC-extra. O novo regionalismo desenvolvimentista sul-americano está finalmente se constituindo para além do policy space do sistema multilateral de comércio, o que, por não configurar, a priori, uma incompatibilidade dos processos de integração com as regras da OMC, contribui para fortalecer o primado do direito internacional na América do Sul. / This thesis aims to answer the following question: Can it be said that the South American developmental regionalism, carried out throughout the twentieth century, and the new developmental regionalism, implemented during the first decade of this century, are compatible with the rules of multilateral trading system? In other words, have the regional integration organizations in South America based their developmental policies in the policy space resulting from the developmental legal prerogatives of the multilateral trading system? To treat such theme, we start from the premise that the countries of South America have historically made use of legal rights for development, since these arose in the multilateral trading system, and enabled the implementation of policies at the national and regional levels. From a legal perspective, the thesis analyses the compatibility of the main developmental aspects of regional integration organizations created in South America since the LAFTA in 1960 to UNASUR, in 2008 with the rules of the multilateral trading system. It was concluded that both the old and the new developmental regionalism in South America were brought forward in accordance with international trade law, either through the GATT period, either during the WTO period. Even with the reduction of policy space resulting from the entry into force of the WTO agreements, South America managed to maintain a relatively high level of policy space for the creation of developmental public policies. Currently, there is a perceived trend of regional integration organizations in South America, especially the MERCOSUR and UNASUR, to carry out public policies on topics unregulated by the WTO, which would characterize a strategy WTO-extra. The new South American developmental regionalism is finally going beyond the policy space of the multilateral trading system, which, by not setting a priori incompatibility of the integration processes with WTO rules, contributes to strengthen the rule of international law in South America.

Thesis (M.A. in Security Studies (Defense Decision-Making and Planning))--Naval Postgraduate School, June 2005. / Thesis Advisor(s): Daniel J. Moran. Includes bibliographical references (p. 95-101). Also available online.

This dissertation analyses the implications of international investment law on host states’ legal ability to protect the environment, regulate sustainable use of natural resources, and develop new approaches to manage environmental risks and uncertainties. ‘Environmental policy space’ is found to be a useful term when exploring the regulatory autonomy in this context. On one hand, investment law aims to ensure stability of the investment environment. On the other hand, environmental law needs flexibility to react to the degradation of the environment. It is found that those different aims do not have to be in conflict. There are useful mechanisms in national environmental law which provide for accessible, transparent and predictable decisions for the private actor. These mechanisms can fulfill the aim of stability in investment law. It is, however, concluded that core provisions of international investment treaties risk to put constraints to environmental law in a variety of ways. To diminish these risks, states, when concluding investment treaties, should make clear that constraining environmental regulation is not compatible with the overarching aim of sustainable development. Furthermore, the interpretation of provisions of investment protection must respect principles and instruments of environmental law not to continue being unbalanced towards investor interests. It is also concluded that allowing for investor – state arbitration, without the investor exhausting local remedies, will ignore the important national administrative review system of public environmental measures.

Submitted by Alice Rocha (rochaalice@yahoo.com.br) on 2012-08-30T01:46:08Z No. of bitstreams: 1 rafael.pdf: 91492 bytes, checksum: b491efbcbf7837043ec36783f3d7f812 (MD5) / Made available in DSpace on 2012-08-30T01:46:08Z (GMT). No. of bitstreams: 1 rafael.pdf: 91492 bytes, checksum: b491efbcbf7837043ec36783f3d7f812 (MD5) / Made available in DSpace on 2013-05-09T21:50:24Z (GMT). No. of bitstreams: 3 rafael.pdf.txt: 7717 bytes, checksum: 953367e7e958b66619c0ac1e8e78694a (MD5) license.txt: 346 bytes, checksum: 6440c47a50909adf871d5cc0caf0b4f9 (MD5) rafael.pdf: 91492 bytes, checksum: b491efbcbf7837043ec36783f3d7f812 (MD5) Previous issue date: 2012-08-29 / A Política de Desenvolvimento Rural que tem sido adotada no Brasil no início do século XXI, que põe em evidência um modelo produtivo baseado na Agricultura Familiar, de pequena-escala, segmento historicamente relegado no País, delineia um caráter inovador na abordagem do Estado brasileiro para com o meio rural nacional. Essa nova abordagem tem propiciado um início de alinhamento do processo de evolução do campo nacional aos objetivos fundamentais da República previstos pela Constituição Federal de 1988, fortalecendo ainda a segurança alimentar da Nação. Por sua vez, o Acordo sobre Agricultura da Organização Mundial do Comércio, na condição de instrumento normativo internacional, no presente contexto de globalização das relações econômicas, apresenta, ao momento de sua renegociação no âmbito da Rodada Doha de negociações multilaterais, um potencial significativo de restringir a capacidade futura do Estado brasileiro de continuar implementando e de aprofundar aquela importante política interna de desenvolvimento, a depender de como os seus termos sejam consolidados. Com isso, configura-se um risco latente de se inviabilizar a própria capacidade nacional de perseguir um modelo de desenvolvimento segundo os ditames da justiça social, solidariedade, e redução da pobreza, da marginalização e das desigualdades, conforme previsto formalmente no texto constitucional. Assim, o presente trabalho buscará mostrar a importância de se preservar, na negociação em curso do novo tratado agrícola multilateral, os importantes espaços de política pública (policy space) necessários para viabilizar a continuidade e o aprofundamento da Política de Desenvolvimento Rural nacional. Para tanto, o presente trabalho buscará: (i) compreender a evolução e situação atual do campo nacional, relacionando-as com a concepção de desenvolvimento prevista pela Constituição Federal brasileira de 1988; (ii) conhecer a Política de Desenvolvimento Rural do País que começou a ser delineada em meados da década de 1990 e foi substancialmente reforçada no período pós-2003, e que tem promovido um redirecionamento do desenvolvimento rural nacional para uma situação mais alinhada à concepção vislumbrada pelos objetivos fundamentais constitucionais, fortalecendo ainda a segurança alimentar doméstica; (iii) em seguida será abordado o Acordo sobre Agricultura da Organização Mundial do Comércio, destacando-se os seus desequilíbrios intrínsecos e as margens diferenciadas de política pública que foram estabelecidas às diferentes categorias de países Membros dentro da Organização, as quais favoreceram os Membros mais desenvolvidos, bem como também será apresentado o contexto da 4 renegociação do tratado agrícola multilateral que está em curso no âmbito da Rodada Doha de negociações; (iv) por fim, serão então analisados e apresentados os potenciais riscos de restrições, diretas e indiretas, que o Acordo sobre Agricultura pode oferecer à capacidade do Estado nacional de continuar e aprofundar a Política de Desenvolvimento Rural do País e a obtenção dos seus resultados esperados.

Unfunded public pension and Other Post Employment Benefits (OPEB) liabilities impose major threats to local fiscal sustainability, which increases governments' default risk and crowds out funding for essential local services. To close the funding gaps, localities may apply a wide range of fiscal instruments, including increasing taxes, fees, and user charges, issuing debt and bonds, obtaining grants and/or decreasing expenditures. This research compares the US local fiscal choice behavior in the context of the fiscal federalism framework. The goal is to identify the ideal mix of constitutional fiscal rules to preserve local fiscal sustainability. Not only should the rules aim to minimize local adverse fiscal behavior pre-crisis, which may include excessive spending, large accumulations of unfunded liabilities, and over-reliance on external grants, but also allow strong local fiscal adaptive capacity post-crisis. The findings help localities identify any effective and prudent fiscal options available to close their pension funding gaps and contribute to the overall sub-national fiscal institutional reforms. Theoretically, this research introduces a novel analytical framework pertaining to local fiscal sustainability by separating pre-crisis and post-crisis institutional analysis and by consolidating two historically viewed as two competing paradigms, public choice and public finance. I argue that the two approaches are complementary rather than contradictory since public choice theory sets up an institutional prerequisite for normative outcomes to be realized and prevents the occurrence of extreme circumstances. The ideal mix of formal fiscal rules, thus, should induce the balanced budget rule that applies to all budget items, stringent spending and debt limits, and institutionalized local tax authority and stable tax structure, but not tax limits. Tax limits are less effective in constraining government than spending and debt limits due to fiscal gimmicks. Moreover, stringent tax limits could significantly limit local governments' ability to bounce back on their own. This research also found that cities do apply different fiscal strategies to reduce exogenous shocks, given their unique fiscal institutions in place. Furthermore, cities with fewer institutional constraints exhibit a faster speed of adjustment. However, certain institutional variables, such as public union size and tax authority, might not have the same fiscal implications as predicted by the theory. Cities often manage to cut their short-term spending regardless of the size of their public unions. A broad range of tax authority does not imply greater local revenue-generating capacity. Own source revenue autonomy might be a better indicator of local fiscal adaptive capacity. / Doctor of Philosophy / Unfunded public pension and Other Post Employment Benefits (OPEB) liabilities impose major threats to local fiscal sustainability, which increases governments’ default risk and crowds out funding for essential local services. To close the funding gaps, localities may apply a wide range of fiscal instruments, including increasing taxes, fees, and user charges, issuing debt and bonds, obtaining grants and/or decreasing expenditures. This research compares the US local fiscal choice behavior in the context of the fiscal federalism framework. The goal is to identify the ideal mix of constitutional fiscal rules to preserve local fiscal sustainability. Not only should the rules aim to minimize local adverse fiscal behavior pre-crisis, which may include excessive spending, large accumulations of unfunded liabilities, and over-reliance on external grants, but also allow strong local fiscal adaptive capacity post-crisis. The findings help localities identify any effective and prudent fiscal options available to close their pension funding gaps and contribute to the overall sub-national fiscal institutional reforms. Theoretically, this research introduces a novel analytical framework pertaining to local fiscal sustainability by separating pre-crisis and post-crisis institutional analysis and by consolidating two historically viewed as two competing paradigms, public choice and public finance. I argue that the two approaches are complementary rather than contradictory since public choice theory sets up an institutional prerequisite for normative outcomes to be realized and prevents the occurrence of extreme circumstances. The ideal mix of formal fiscal rules, thus, should induce the balanced budget rule that applies to all budget items, stringent spending and debt limits, and institutionalized local tax authority and stable tax structure, but not tax limits. Tax limits are less effective in constraining government than spending and debt limits due to fiscal gimmicks. Moreover, stringent tax limits could significantly limit local governments’ ability to bounce back on their own. This research also found that cities do apply different fiscal strategies to reduce exogenous shocks, given their unique fiscal institutions in place. Furthermore, cities with fewer institutional constraints exhibit a faster speed of adjustment. However, certain institutional variables, such as public union size and tax authority, might not have the same fiscal implications as predicted by the theory. Cities often manage to cut their short-term spending regardless of the size of their public unions. A broad range of tax authority does not imply greater local revenue-generating capacity. Own source revenue autonomy might be a better indicator of local fiscal adaptive capacity.

For the most part, formal economic integration between African states can be characterised as ‘theatre regionalism’: governments sign regional economic agreements with no intention to implement them. Yet amidst widespread theatre there have been a few instances of actual integration. This thesis sets out to explain this variance: under what conditions do African governments implement – and not just sign – formal agreements on regional economic integration? To answer this question the dominant Eurocentric literature on comparative regionalism is amended with insights from the third worldist literature on African states to develop a new approach for comparative analysis, the ‘Regionalism as Policy Space’ (RPS) framework. This framework models African regionalism as a two-stage game. At the first stage, governments’ interests in regionalism are determined by perceptions of the existence of structural cross-issue linkages connecting implementation of regional agreements with the widening of government policy space. Given such linkages, at the second stage, governments of a region engage in a coordination game to establish the distribution of benefits from integration. Variance in the implementation of regional agreements, then, is explained by variance in the existence of perceived cross-issue linkages (the Benefits Existence Condition) and the ability of participating governments to ease distributional tensions (the Benefits Distribution Condition). Four African customs union case studies - the East African customs union of the 1960s and 70s, the customs union of the East African Community in the 2000s, the customs union of the Economic Community of West African States and the Southern African Customs Union – lend strong empirical support to the RPS framework. The thesis ends with a discussion of the role of hegemons and proposes a series of policy measures aiming to reduce the likelihood of theatre regionalism in Africa.

Magister Legum - LLM

No description available.