Quantum-Inspired Portfolio Optimization In The QUBO Framework

TL;DR

This paper introduces a quantum-inspired optimization method for portfolio selection that improves computational efficiency and solution accuracy, validated on real-world financial data and enhanced by a novel preprocessing technique.

Contribution

It presents a new quantum-inspired approach combined with a two-stage search preprocessing to optimize asset portfolios more effectively than traditional methods.

Findings

Faster and more accurate portfolio optimization results.

Effective enhancement of computational efficiency with parameter tuning.

Demonstrated potential of quantum-inspired techniques in financial asset management.

Abstract

A quantum-inspired optimization approach is proposed to study the portfolio optimization aimed at selecting an optimal mix of assets based on the risk-return trade-off to achieve the desired goal in investment. By integrating conventional approaches with quantum-inspired methods for penalty coefficient estimation, this approach enables faster and accurate solutions to portfolio optimization which is validated through experiments using a real-world dataset of quarterly financial data spanning over ten-year period. In addition, the proposed preprocessing method of two-stage search further enhances the effectiveness of our approach, showing the ability to improve computational efficiency while maintaining solution accuracy through appropriate setting of parameters. This research contributes to the growing body of literature on quantum-inspired techniques in finance, demonstrating its potential as a useful tool for asset allocation and portfolio management.