Heuristic Optimization of Robust Portfolios Under Stochastic Environments

In this paper I show how a heuristic optimization algorithm can be used to optimize robust portfolios of assets with arbitrary distributions. To generate asset returns I introduce a Monte Carlo simulation framework that is based on stochastic stock and interest rate models. Robust portfolio selection results in portfolios that deliver stable returns under various market scenarios at the same time. I formulate the optimization problem for several risk and performance measures and adapt a simulated annealing algorithm to solve the complex optimization problem. The presented concepts are implemented in a Matlab/Excel application which is used to optimize a real life robust portfolio. I show the feasibility of the approach and analyze the benefits of robust portfolios.