Portfolio Optimization with Classical and Quantum Methods

This repository implements classical portfolio optimization using the Markowitz mean–variance framework and establishes a strong baseline for future quantum optimization using QUBO and QAOA.

The current focus is on data preparation, convex optimization with modern solvers, and empirical analysis on real market data.

---

What’s Implemented

• Markowitz Mean–Variance Portfolio Optimization
• Classical optimization using CVXPY with the Clarabel solver
• Monte Carlo portfolio simulation
• Real market data preprocessing (NIFTY 50)
• Sensitivity analysis across different risk preferences

---

Work Completed

1. Theory & Background

• Studied Modern Portfolio Theory and classical portfolio optimization methods
• Reviewed optimization formulations relevant to quantum computing
• Authored a 3–4 page literature review on classical and emerging quantum approaches

2. Classical Optimization Baseline

• Selected 10 stocks across 3 different sectors
• Computed optimized portfolios for three risk-aversion levels
• Compared:
  • Monte Carlo–simulated portfolios
  • Convex optimization solutions using Clarabel
• Analyzed expected return, variance, and asset allocation behavior

3. Data Engineering

• Dataset: NIFTY 50 stock market data
• Performed data cleansing, consolidation, and yearly return computation
• Built reusable preprocessing pipelines for optimization experiments

4. Solver Analysis

• Implemented optimization solvers in Python
• Evaluated:
  • Optimized portfolio weights
  • Expected return and risk
• Re-estimated portfolios under different solver parameters and interpreted asset selection

---

Tools & Technologies

• Python
• NumPy, Pandas
• CVXPY (Clarabel solver)
• SciPy
• Matplotlib
• Jupyter Notebook

---