Local Search Techniques for Constrained Portfolio Selection Problems

We consider the problem of selecting a portfolio of assets that provides theinvestor a suitable balance of expected return and risk. With respect to theseminal mean-variance model of Markowitz, we consider additionalconstraints on the cardinality of the portfolio and on the quantity ofindividual shares. Such constraints better capture the real-world tradingsystem, but make the problem more difficult to be solved with exact methods.We explore the use of local search techniques, mainly tabu search, for theportfolio selection problem. We compare the combine previous work on portfolioselection that makes use of the local search approach and we propose newalgorithms that combine different neighborhood relations. In addition, we showhow the use of randomization and of a simple form of adaptiveness simplifiesthe setting of a large number of critical parameters. Finally, we show how ourtechniques perform on public benchmarks.