Project Overview

This project explores portfolio optimization techniques for constructing efficient portfolios under long-only constraints. Using modern portfolio theory and numerical optimization, we analyze risk-return trade-offs in real-world investing scenarios where short selling is not permitted.

The analysis combines Monte Carlo simulation, Markowitz mean-variance optimization, and Sequential Least Squares Programming (SLSQP) to construct efficient frontiers under various practical constraints including minimum/maximum weights and sector limitations.

💰 Key Finance Concepts

Efficient Frontier: The set of optimal portfolios offering the highest expected return for each level of risk.

Sharpe Ratio: A measure of risk-adjusted return, calculated as excess return per unit of volatility.

Capital Market Line (CML): The line connecting the risk-free asset to the market portfolio, representing optimal risk-return combinations.

Long-Only Constraints: Investment restrictions that prevent short selling, requiring all portfolio weights to be non-negative.

🎯 Project Objectives

  • Implement Markowitz mean-variance optimization
  • Apply real-world long-only investment constraints
  • Compare Monte Carlo vs. analytical optimization
  • Analyze the impact of weight constraints on efficiency
  • Visualize efficient frontiers and optimal allocations

🔧 Optimization Methods

  • Monte Carlo simulation of random portfolios
  • Sequential Least Squares Programming (SLSQP)
  • Constrained quadratic programming
  • Sharpe ratio maximization
  • Minimum variance portfolio construction

📊 Analysis Features

  • Efficient frontier construction and visualization
  • Portfolio allocation pie charts
  • Risk-return scatter plots
  • Constraint impact analysis
  • Capital Market Line identification

📈 Key Visualizations

Efficient frontier comparison between SLSQP optimization and random portfolios

SLSQP Optimization vs Random Portfolios

Impact of minimum weight constraints on efficient frontier

Impact of Minimum Weight Constraints

⚠️ Current Limitations

  • Based on historical mean returns and covariances
  • Assumes normally distributed returns
  • No transaction costs or turnover constraints
  • Static optimization (single-period model)
  • Limited to publicly available market data