Abstratct: This study employs an ARMA-GARCH-EVT modeling approach to capture marginal features and Vine copula models to understand tail dependence in a portfolio of 60 stocks listed on the Brazilian stock market, specifically the Ibovespa. Throughout this analysis, we focus on portfolio optimization using five key investment performance metrics, including Information Ratio, Sharpe Ratio, and Sortino Ratio. These optimizations were performed at two rebalancing frequencies, all focusing on risk mitigation and index optimization, and considering market transaction costs. Notably, one of the significant findings of this study is the ability to achieve positive excess returns compared to the Ibovespa Index during 2019 and 2020. It is worth noting that this superior performance is particularly worthy of note, as it is observed, in terms of accumulated return, in 17 of the 20 portfolios analyzed
Palavras-Chave: Portfolio optimization, Vine-Copula, Extreme value theory, Ibovespa.