Ensaios sobre mercados artificiais com dois ativos utilizando algoritmos genéticos

• Autor(a) principal: Monch, Erich
• Data de Publicação: 2016
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional do FGV (FGV Repositório Digital)
• Texto Completo: http://hdl.handle.net/10438/15513
• topic: Preços; Mercado; Economia; Finanças; Mercado financeiro; Algoritmos genéticos; Métodos de simulação

Agent-Based Modelling is an approach for complex systems used to search for answers beginning with the analisis and construction of small components and their interactions. The observed results of the simulations are aggregated and derive from the combination of actions e interferences occurring in the microscopic level of the model. Thereby, leads to a bottom-up simulation. By meeting all the requirements, Financial Markets are perfect systems to use these models. This work implements an Agent-Based Financial Market Model composed by diverse agents interacting by themselves through a Dealing Center to trade two assets and counts on the service of Market Makers to promote market liquidity as seen in real markets. For this model to work, two types of agents that manage portfolios with both assets simultaneously are developed. The first one uses the Markowitz model meanwhile the second one uses Asset-Spread analysis techniques. Another contribution from this model is the analysis about the Objective Function on Assets Returns in the place of analysis on Prices.