Análise de volatilidade, integração de preços e previsibilidade para o mercado brasileiro de camarão

The present paper has the purpose of investigate the dynamics of the volatility structure in the shrimp prices in the Brazilian fish market. Therefore, a description of the initial aspects of the shrimp price series was made. From this information, statistics tests were made and selected univariate models to be price predictors. Then, it was verified the existence of relationship of long-term equilibrium between the Brazilian and American imported shrimp and if, confirmed the relationship, whether or not there is a causal link between these assets, considering that the two countries had presented trade relations over the years. It is presented as an exploratory research of applied nature with quantitative approach. The database was collected through direct contact with the Companhia de Entrepostos e Armazéns Gerais de São Paulo (CEAGESP) and on the official website of American import, National Marine Fisheries Service - National Oceanic and Atmospheric Administration (NMFS- NOAA). The results showed that the great variability in the active price is directly related with the gain and loss of the market agents. The price series presents a strong seasonal and biannual effect. The average structure of price of shrimp in the last 12 years was R$ 11.58 and external factors besides the production and marketing (U.S. antidumping, floods and pathologies) strongly affected the prices. Among the tested models for predicting prices of shrimp, four were selected, which through the prediction methodologies of one step forward of horizon 12, proved to be statistically more robust. It was found that there is weak evidence of long-term equilibrium between the Brazilian and American shrimp, where equivalently, was not found a causal link between them. We concluded that the dynamic pricing of commodity shrimp is strongly influenced by external productive factors and that these phenomena cause seasonal effects in the prices. There is no relationship of long-term stability between the Brazilian and American shrimp prices, but it is known that Brazil imports USA production inputs, which somehow shows some dependence productive. To the market agents, the risk of interferences of the external prices cointegrated to Brazilian is practically inexistent. Through statistical modeling is possible to minimize the risk and uncertainty embedded in the fish market, thus, the sales and marketing strategies for the Brazilian shrimp can be consolidated and widespread

Aprendizado por Reforço com Valores deInfluência em Sistemas Multi-Agente

We propose a new paradigm for collective learning in multi-agent systems (MAS) as a solution to the problem in which several agents acting over the same environment must learn how to perform tasks, simultaneously, based on feedbacks given by each one of the other agents. We introduce the proposed paradigm in the form of a reinforcement learning algorithm, nominating it as reinforcement learning with influence values. While learning by rewards, each agent evaluates the relation between the current state and/or action executed at this state (actual believe) together with the reward obtained after all agents that are interacting perform their actions. The reward is a result of the interference of others. The agent considers the opinions of all its colleagues in order to attempt to change the values of its states and/or actions. The idea is that the system, as a whole, must reach an equilibrium, where all agents get satisfied with the obtained results. This means that the values of the state/actions pairs match the reward obtained by each agent. This dynamical way of setting the values for states and/or actions makes this new reinforcement learning paradigm the first to include, naturally, the fact that the presence of other agents in the environment turns it a dynamical model. As a direct result, we implicitly include the internal state, the actions and the rewards obtained by all the other agents in the internal state of each agent. This makes our proposal the first complete solution to the conceptual problem that rises when applying reinforcement learning in multi-agent systems, which is caused by the difference existent between the environment and agent models. With basis on the proposed model, we create the IVQ-learning algorithm that is exhaustive tested in repetitive games with two, three and four agents and in stochastic games that need cooperation and in games that need collaboration. This algorithm shows to be a good option for obtaining solutions that guarantee convergence to the Nash optimum equilibrium in cooperative problems. Experiments performed clear shows that the proposed paradigm is theoretical and experimentally superior to the traditional approaches. Yet, with the creation of this new paradigm the set of reinforcement learning applications in MAS grows up. That is, besides the possibility of applying the algorithm in traditional learning problems in MAS, as for example coordination of tasks in multi-robot systems, it is possible to apply reinforcement learning in problems that are essentially collaborative