Investment portfolio tracking using model predictive control
Carregando...
Data
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Minas Gerais
Descrição
Tipo
Artigo de periódico
Título alternativo
Primeiro orientador
Membros da banca
Resumo
The composition of an investment portfolio aiming to increase the financial returns while reducing the exposure to risk is a topic of growing interest in the world. In this direction, we propose a model predictive control (MPC) strategy in order to optimize the investment portfolio selection taking into account the tracking of a reference portfolio with desired return. In addition, an analysis comparing different sizes of the prediction horizon according to VPH‐MPC (Varying Predictive Horizon‐MPC) and FPH‐MPC (Fixed Predictive Horizon‐MPC) algorithms is conducted. Finally, we propose an optimal control problem using the tracking error as a function of loss of CVaR (Conditional Value at Risk) measurement. Numerical experiments are run based on Brazilian Stock Exchange data. The experimental results are compared with the Markowitz portfolio optimization model, a conventional tracking strategy, and benchmarks from the Brazilian financial market. This comparison indicates a good tracking performance obtained by the proposed MPC in the two versions while satisfying the constraints.
Abstract
Assunto
Teoria do controle
Palavras-chave
model predictive control, optimal control, portfolio tracking and selection
Citação
Departamento
Curso
Endereço externo
https://onlinelibrary.wiley.com/doi/10.1002/oca.2937