Using Quantum Generative Adversarial Networks for Portfolio Optimisation

Calum Holker, Pavan Jayasinha, Aarsh Chaube, Tomasz Andrzejewski
Submission for QHack open hackathon for team QLords

Blog Post: Using Quantum Generative Adversarial Networks for Portfolio Optimisation

Project Summary

We use qGANs on stock market data to create a model that can predict future trends, and look at how that data can then be used to optimise portfolios. More details on both of these implementations can be found within the READMEs in their respective directores

Built With

PennyLane
Qiskit
Amazon Braket
Google FLOQ

File Structure

data/ - contains stock market data in sequences, and notbooks creating that data
qGAN/ - contains the implementation of qGANs on stock market data
QAOA-VQE/ - contains the implementation of QAOA and VQE for portfolio optimisation
testing-versions/ - contains files used for testing the main files

References

(1) PAGAN: Portfolio Analysis with Generative Adversarial Networks
(2) Quantum Generative Adversarial Networks for learning and loading random distributions
(3) Enhancing Combinatorial Optimization with Quantum Generative Models
(4) Improving Variational Quantum Optimization using CVaR
(5) Qiskit Aqua Portfolio Optimisation