Fast Portfolio Diversification

Abstract

This is a concise description of a new algorithm for diversifying an investment portfolio, that was previously described and tested in great detail in another paper, which also showed that the algorithm is very fast and guaranteed to converge to the optimal solution in just a few iterations, and the algorithm improves the portfolio on several performance metrics, while being extremely robust to estimation errors in the correlation matrix. This paper further shows how to implement the algorithm for a sparse correlation matrix, and it compares the time-usage of the sparse and dense versions of the algorithm, as well as parallel and serial versions.

Keywords: investing, portfolio optimization, diversification