Portfolio Optimization: An Operations Research Approach

Event Type

Research Presentation

Academic Department

Mathematics and Statistics

Location

Dana Science Building, 2nd floor

Start Date

24-4-2026 1:00 PM

End Date

24-4-2026 2:30 PM

Description

This thesis applies operations research methods to financial portfolio optimization using rigorous models derived from convex optimization and linear programming. Three models are developed and compared: Minimum Variance, Maximum Sharpe Ratio (Charnes-Cooper transformation), and CVaR Minimization (Rockafellar-Uryasev LP). Each model is implemented in Python using CVXPY and validated against equal-weight and risk-parity benchmarks via walk-forward backtesting. A six-week live trading simulation on the Alpaca Markets paper trading API tests whether theoretical allocations can be faithfully executed in practice. Central finding: Formal optimization offers measurable advantages over naive diversification under backtesting conditions, while the live trading window reveals meaningful divergence between theoretical allocations and execution.

Comments

Under the direction of Dr. Giancarlo Schrementi.

This document is currently not available here.

Share

COinS
 
Apr 24th, 1:00 PM Apr 24th, 2:30 PM

Portfolio Optimization: An Operations Research Approach

Dana Science Building, 2nd floor

This thesis applies operations research methods to financial portfolio optimization using rigorous models derived from convex optimization and linear programming. Three models are developed and compared: Minimum Variance, Maximum Sharpe Ratio (Charnes-Cooper transformation), and CVaR Minimization (Rockafellar-Uryasev LP). Each model is implemented in Python using CVXPY and validated against equal-weight and risk-parity benchmarks via walk-forward backtesting. A six-week live trading simulation on the Alpaca Markets paper trading API tests whether theoretical allocations can be faithfully executed in practice. Central finding: Formal optimization offers measurable advantages over naive diversification under backtesting conditions, while the live trading window reveals meaningful divergence between theoretical allocations and execution.