Portfolio Dynamic Optimisation based on GARCH Volatility Forecasting

Abstract

This study focuses on the portfolio optimisation problem in the financial market, aiming to construct a portfolio model that integrates GARCH volatility prediction and dynamic optimisation. By collecting daily frequency data of five weighted stocks in the S&P 500 constituents, this article use GARCH (1,1) and EGARCH models to predict asset volatility and design a rebalancing strategy by combining dynamic programming and quadratic programming. The study addresses the key issues of coupling the GARCH prediction results with the dynamic optimisation model, optimising the efficiency of quadratic programming under high-dimensional data, and verifying the stability of the dynamic strategy. The results show that the constructed model outperforms the static strategy in terms of return-risk performance, provides investors with more effective risk management tools, and enriches the theory and practice of financial risk management and asset allocation.