Title: Algorithmic Trading Using Genetic Algorithms in the Brazilian Stock Exchange
Authors: Jean Pierre Jarrier Conti, Heitor Silvério Lopes
Abstract: The evolution of the computational capacity has been helping financial markets to increase the success in their operational running strategies on its investment portfolios. After stock market evolved to make all its operations electronically a new approach called algorithmic trading has gained attention from academic researches. This paper presents a novel method of the dynamic optimization to improve the profit of the algorithmic trading. Combining two genetic algorithms, the proposed approach seek to finding the best optimization and trading window for a trading strategy. The performance of this approach was evaluated with data of the last five years of two stocks traded at the Brazilian Stock Exchange. Comparing the results obtained with classical moving averages indicators, the proposed method performed better in all cases using the complete dataset and using year by year, all experiments using shares of PETR4. These results suggest that the discovery of the optimal trading and optimization window we can improve the system trading strategy and lead to increased profits.
Key-words: Finance, optimization, evolutionary computation, genetic algorithm, algorithmic trading
DOI code: 10.21528/CBIC2019-112
PDF file: CBIC2019-112.pdf
BibTeX file: CBIC2019-112.bib