Comparing Classical-Quantum Portfolio Optimization with Enhanced Constraints

Salvatore Certo,Anh Dung Pham,Daniel Beaulieu

DOI: https://doi.org/10.48550/arXiv.2203.04912

2022-03-10

Abstract:One of the problems frequently mentioned as a candidate for quantum advantage is that of selecting a portfolio of financial assets to maximize returns while minimizing risk. In this paper we formulate several real-world constraints for use in a Quantum Annealer (QA), extending the scenarios in which the algorithm can be implemented. Specifically, we show how to add fundamental analysis to the portfolio optimization problem, adding in asset-specific and global constraints based on chosen balance sheet metrics. We also expand on previous work in improving the constraint to enforce investment bands in sectors and limiting the number of assets to invest in, creating a robust and flexible solution amenable to QA. Importantly, we analyze the current state-of-the-art algorithms for solving such a problem using D-Wave's Quantum Processor and compare the quality of the solutions obtained to commercially-available optimization software. We explore a variety of traditional and new constraints that make the problem computationally harder to solve and show that even with these additional constraints, classical algorithms outperform current hybrid solutions in the static portfolio optimization model.

Materials Science,Quantum Physics

What problem does this paper attempt to address?

This paper aims to explore how to implement the portfolio optimization problem with enhanced constraints on a quantum annealer (QA) and compare it with classical optimization algorithms. Specifically, the paper focuses on how to add various real - world constraints in portfolio selection, such as fundamental analysis based on balance - sheet indicators, industry investment proportion limits, and limits on the number of investment assets, etc., in order to construct more robust and flexible portfolio optimization solutions.

### Main problems solved in the paper:

1. **Introducing practical constraints**: The paper proposes several methods for adding constraints in the actual portfolio optimization during the quantum annealing process, such as asset - specific and global constraints, which are based on selected balance - sheet indicators.

2. **Expanding constraints**: In addition to the basic portfolio optimization problem, the paper also expands the constraints, including:
- **Industry investment proportion limits**: By introducing slack variables to set the maximum and minimum investment proportions for each industry.
- **Asset number limits**: Limit the number of assets included in the portfolio, which is very important in actual investment, because factors such as trading costs and monitoring overhead usually require investors to limit the number of selected assets.
- **Fundamental analysis**: Construct a portfolio according to the company's financial indicators (such as the current ratio) to ensure that the portfolio not only maximizes returns but also has specific characteristics desired by investors.

3. **Performance comparison**: The paper uses D - Wave's quantum processor and commercially available classical optimization software (such as CPLEX) to solve these optimization problems and compares the performance of different algorithms in dealing with portfolio optimization problems with complex constraints.

### Main conclusions:

- **Better performance of classical algorithms**: Even with the addition of multiple complex real - world constraints, classical optimization algorithms (such as CPLEX) are still superior to the current hybrid quantum solutions in the static portfolio optimization model.
- **Advantages of quantum annealing**: Although currently quantum annealers are still limited by hardware when dealing with large - scale problems, with the development of quantum hardware, quantum annealing may show advantages in dealing with certain types of problems.
- **Flexibility and practicality**: The methods proposed in the paper provide more flexibility and practicality for portfolio optimization in practical applications, especially when multiple complex constraints need to be considered.

Through these studies, the paper provides an important reference and basis for the future application of quantum computing in the financial field.