Modelo heterogêneo adaptativo com informação para a dinâmica da distribuição de renda

• Autor(a) principal: COSTA NETO, Miguel Alves da
• Data de Publicação: 2019
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UFRPE
• Texto Completo: http://www.tede2.ufrpe.br:8080/tede2/handle/tede2/8522
• topic: Distribuição de renda; Modelo computacional; Informação; CIENCIAS EXATAS E DA TERRA::FISICA

In the last four decades, physicists have investigated economic problems from as their experiences in creating mathematical models to describe natural phenomena. Among these, there can be highlighted the studies related to the characterization of the inequalities of the income distribution associated with the agent-based models. One way of quantifying such inequalities occurs through the calculation of the Gini index 0 G 1, a parameter which indicates maximum inequality for G = 1 and minimum when G = 0. In this master thesis we present a computational model, called adaptive heterogeneous model with information, in it a fraction q of its agents has a negative correlation between its expenditure !i and its income level mi, while the remaining population has a positive correlation. We analyzed the role of asymmetry of information on distribution through the parameter p, which establishes the fraction of agents with global information of the average value of the transactions. We investigated the polarization of the distribution by means of the bimodality coefficient and study its dependence as a function of the parameters p and q. The model produces several inequality, polarization, and resource flows as emerging characteristics of dynamics, in particular the states (G; !) corresponds to limiting behaviors compatible with those observed in real data. In general, distributions exhibit a lower bimodal feature for situations of greater uniformity independent on adaptation fraction p. For the intermediate situations (0 < q < 1) polarization presents a monotonically decreasing behavior with p.