Global ETD Search

1	Uso de política abstrata estocástica na navegação robótica. / Using stochastic abstract policies in robotic navigation. Matos, Tiago 06 September 2011 (has links) A maioria das propostas de planejamento de rotas para robôs móveis não leva em conta a existência de soluções para problemas similares ao aprender a política para resolver um novo problema; e devido a isto, o problema de navegação corrente deve ser resolvido a partir do zero, o que pode ser excessivamente custoso em relação ao tempo. Neste trabalho é realizado o acoplamento do conhecimento prévio obtido de soluções similares, representado por uma política abstrata, a um processo de aprendizado por reforço. Além disto, este trabalho apresenta uma arquitetura para o aprendizado por reforço simultâneo, de nome ASAR, onde a política abstrata auxilia na inicialização da política para o problema concreto, e ambas as políticas são refinadas através da exploração. A fim de reduzir a perda de informação na construção da política abstrata é proposto um algoritmo, nomeado X-TILDE, que constrói uma política abstrata estocástica. A arquitetura proposta é comparada com um algoritmo de aprendizado padrão e os resultados demonstram que ela é eficaz em acelerar a construção da política para problemas práticos. / Most work in path-planning approaches for mobile robots does not take into account existing solutions to similar problems when learning a policy to solve a new problem, and consequently solves the current navigation problem from scratch, what can be very time consuming. In this work we couple a prior knowledge obtained from a similar solution to a reinforcement learning process. The prior knowledge is represented by an abstract policy. In addition, this work presents a framework for simultaneous reinforcement learning called ASAR, where the abstract policy helps start up the policy for the concrete problem, and both policies are refined through exploration. For the construction of the abstract policy we propose an algorithm called X-TILDE, that builds a stochastic abstract policy, in order to reduce the loss of information. The proposed framework is compared with a default learning algorithm and the results show that it is effective in speeding up policy construction for practical problems. Aprendizado computacional relacional Inteligência artificial Knowledge transfer Policy abstraction Reinforcement learning Relational MDP Robotic navigation Robótica
2	Uso de política abstrata estocástica na navegação robótica. / Using stochastic abstract policies in robotic navigation. Tiago Matos 06 September 2011 (has links) A maioria das propostas de planejamento de rotas para robôs móveis não leva em conta a existência de soluções para problemas similares ao aprender a política para resolver um novo problema; e devido a isto, o problema de navegação corrente deve ser resolvido a partir do zero, o que pode ser excessivamente custoso em relação ao tempo. Neste trabalho é realizado o acoplamento do conhecimento prévio obtido de soluções similares, representado por uma política abstrata, a um processo de aprendizado por reforço. Além disto, este trabalho apresenta uma arquitetura para o aprendizado por reforço simultâneo, de nome ASAR, onde a política abstrata auxilia na inicialização da política para o problema concreto, e ambas as políticas são refinadas através da exploração. A fim de reduzir a perda de informação na construção da política abstrata é proposto um algoritmo, nomeado X-TILDE, que constrói uma política abstrata estocástica. A arquitetura proposta é comparada com um algoritmo de aprendizado padrão e os resultados demonstram que ela é eficaz em acelerar a construção da política para problemas práticos. / Most work in path-planning approaches for mobile robots does not take into account existing solutions to similar problems when learning a policy to solve a new problem, and consequently solves the current navigation problem from scratch, what can be very time consuming. In this work we couple a prior knowledge obtained from a similar solution to a reinforcement learning process. The prior knowledge is represented by an abstract policy. In addition, this work presents a framework for simultaneous reinforcement learning called ASAR, where the abstract policy helps start up the policy for the concrete problem, and both policies are refined through exploration. For the construction of the abstract policy we propose an algorithm called X-TILDE, that builds a stochastic abstract policy, in order to reduce the loss of information. The proposed framework is compared with a default learning algorithm and the results show that it is effective in speeding up policy construction for practical problems. Aprendizado computacional relacional Inteligência artificial Robótica Knowledge transfer Policy abstraction Reinforcement learning Relational MDP Robotic navigation
3	FÖRESTÄLLNINGAR OCH INTRESSEN : En fallstudie utifrån Advocacy Coalition Framework av en lokal policyprocess om expropriation Blomqvist, Fredrik January 2016 (has links) This paper examines the viability of the Advocacy Coalition Framework(ACF) by applying it in a single case study. The aim is to advance the framework’s theoretical understanding of the policy process and its usefulness for analyzing local policy contexts. The case addressed is a long-spun policy conflict regarding the use of compulsory acquisition of real estate by a Swedish municipality for the sake of local business development. Analyzed data consisted of the municipality diary on the issue, correspondence between actors, public statements, official and internal documents and interviews with actors and non-actors. The ACF is a good starting point for understanding this local policy process, largely because of the great flexibility of its concepts. However, its basic assumptions on beliefs cannot fully explain observed events. Relating to this, the paper has five main findings. First, although beliefs play an important role in forming policy action, so does interests. Second, a conjunction of beliefs and self-interest is an important condition for some actors’ actions. Third, coalition formation is not dependent on similarity of beliefs but on similarity of policy objectives. Fourth, policy objectives are resultant of beliefs for some actors, of self-interest for others and for yet others the result of both. Therefore, actors in coalition act to achieve the same policy objectives but not necessarily for the same reasons. Fifth, one non-actor refrained from policy action in spite of strong policy core beliefs due to the policy process not being a zero sum game for this non-actor. This paper supports recent studies proposing the incorporation of interests into the ACF. For further development of the ACF the paper suggests further research to answer two generic questions: What is the relationship betweeninterests and beliefs? Are potential actors more likely to take policy action inzero sum game policy processes? For the ACF to cope with certain conditionsin local contexts the paper suggest further research into the question: Is the level of abstraction of policy issues key in understanding the involvement of legal and natural persons and their basis for policy action? policy process policyprocess advocacy coalition framework advocacy coalition framework acf beliefs interests beliefs and interests interests and beliefs local policy municipality sweden scandinavia compulsory acquisition of real estate rule of law business trade industry business development coalition formation action policy action policy objectives objectives zero sum game policy abstraction realism constructivism policy process policyprocess advocacy coalition framework advocacy coalition framework acf påverkan koalition föreställningar intressen värderingar beliefs interests sverige realism constructivism lokal politik local policy kommunpolitik kommun näringsliv expropriation rättssäkerhet rättsstat fastighet plan och bygglagen näringslivsutveckling näringsliv koalitionsformering policyagerande action policymål objectives nollsummespel zero sum game policyabstraktion policyabstraction

1

Page generated in 0.0925 seconds