quantum-portfolio-opt 0.1.0

Hybrid Quantum-Classical Portfolio Optimization with IBM Qiskit

Quantfolio 🚀⚛️

Hybrid Quantum-Classical Portfolio Optimization powered by IBM Qiskit and Deep Learning.

Quantfolio combines classical machine learning (Transformers/LSTM) with quantum optimization algorithms (QAOA) to optimize financial portfolios, offering a cutting-edge approach to asset allocation.

---

✨ Features

• 🌌 Quantum Optimization: QAOA (Quantum Approximate Optimization Algorithm) for portfolio selection
• 🧠 Classical ML: Transformer and LSTM models for return prediction
• 📊 Hybrid Approach: Best of both worlds - ML predictions + quantum optimization
• 📈 Financial Metrics: Sharpe Ratio, Sortino Ratio, VaR, CVaR, Maximum Drawdown
• 🎨 Visualization: Beautiful plots for weights, efficient frontier, and portfolio evolution
• ⚡ Easy API: Simple interface for quick prototyping and production use

---

🔧 Installation

pip install quantum-portfolio-opt

Requirements: Python 3.9+

---

🚀 Quick Start

Simple Optimization

import quantfolio as qf

# Optimize portfolio with quantum algorithm
optimizer = qf.HybridOptimizer(
    tickers=['AAPL', 'MSFT', 'GOOGL', 'AMZN'],
    method='quantum'
)

weights = optimizer.optimize(risk_aversion=0.5)

print("Optimal Portfolio:")
for ticker, weight in zip(optimizer.tickers, weights):
    print(f"  {ticker}: {weight*100:.1f}%")

Advanced Usage

import quantfolio as qf

# Hybrid optimization with ML prediction
optimizer = qf.HybridOptimizer(
    tickers=['AAPL', 'MSFT', 'GOOGL', 'AMZN', 'JPM', 'JNJ'],
    classical_model='transformer',  # or 'lstm'
    use_quantum=True,
    risk_aversion=0.5,
    max_weight=0.30  # Max 30% per asset
)

# Train predictor and optimize
result = optimizer.optimize(
    train_predictor=True,
    epochs=50
)

# Visualize results
optimizer.plot_weights()
optimizer.plot_efficient_frontier()

# View metrics
print(f"Expected Return: {result['expected_return']*100:.2f}%")
print(f"Risk (Volatility): {result['risk']*100:.2f}%")
print(f"Sharpe Ratio: {result['sharpe_ratio']:.2f}")

Compare Methods

import quantfolio as qf

# Compare quantum vs classical vs hybrid
comparison = qf.compare_methods(
    tickers=['AAPL', 'MSFT', 'GOOGL'],
    methods=['quantum', 'classical', 'hybrid']
)

print(comparison)

---

📊 Available Optimizers

Optimizer	Description	Use Case
QAOA	Quantum Approximate Optimization	Complex constraints, exploration
Markowitz	Mean-Variance Optimization	Classical baseline, proven
Equal Weight	1/N allocation	Simple diversification
Min Variance	Minimize risk only	Conservative portfolios
Max Sharpe	Maximize risk-adjusted return	Optimal risk/return

---

🎓 Examples

Check out the examples/ directory for:

• basic_usage.py - Quick start tutorial
• custom_portfolio.py - Advanced configurations
• comparison_demo.py - Method comparisons
• backtesting.py - Historical performance testing

---

📖 Documentation

Main API

HybridOptimizer

optimizer = qf.HybridOptimizer(
    tickers: List[str],              # Stock symbols
    method: str = 'hybrid',          # 'quantum', 'classical', or 'hybrid'
    classical_model: str = 'lstm',   # 'transformer' or 'lstm'
    use_quantum: bool = True,        # Use QAOA optimization
    risk_aversion: float = 0.5,      # 0=aggressive, 1=conservative
    max_weight: float = 0.30,        # Maximum weight per asset
    min_weight: float = 0.05         # Minimum weight per asset
)

Methods

• optimize(**kwargs) - Run optimization
• plot_weights() - Visualize portfolio weights
• plot_efficient_frontier() - Plot risk-return trade-off
• plot_evolution() - Show portfolio value over time

---

🧪 Technical Details

Quantum Component

• Algorithm: QAOA (Quantum Approximate Optimization Algorithm)
• Backend: IBM Qiskit Aer Simulator (portable to real quantum hardware)
• Qubits: Configurable based on number of assets
• Encoding: Binary representation of portfolio weights

Classical Component

• Transformers: Multi-head attention for time series
• LSTM: Bidirectional LSTM for sequence modeling
• Features: Price history, technical indicators
• Framework: PyTorch

Optimization Problem

Maximizes: Expected Return - λ * Risk

Subject to:

• Σ weights = 1 (fully invested)
• 0 ≤ weight_i ≤ max_weight
• Minimum diversification constraints

---

⚠️ Disclaimer

This software is for educational and research purposes only.

• Not financial advice
• No guarantee of returns
• Past performance ≠ future results
• Consult a qualified financial advisor before investing

---

🤝 Contributing

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.

Development Setup

git clone https://github.com/quantfolio/quantfolio.git
cd quantfolio
pip install -e ".[dev]"
pytest tests/

---

📚 Citation

If you use Quantfolio in academic research, please cite:

@software{quantfolio2025,
  title = {Quantfolio: Hybrid Quantum-Classical Portfolio Optimization},
  author = {Quantfolio Contributors},
  year = {2025},
  url = {https://github.com/quantfolio/quantfolio}
}

---

📄 License

MIT License - see LICENSE file for details.

---

🌟 Acknowledgments

Built with:

• IBM Qiskit - Quantum computing framework
• PyTorch - Deep learning
• yfinance - Financial data

---

📬 Contact

• Issues: GitHub Issues
• Discussions: GitHub Discussions

---

Made with ❤️ and ⚛️ by the Quantfolio team