Money in trees: How memes, trees, and isolation can optimize financial portfolios

In this paper, we propose the hybrid application of two nature inspired approaches to the problem of Portfolio Optimization. This problem consists of the selection and weighting of financial assets. Its goal is to form an investment strategy which maximizes a return measure and minimizes a risk measure. We perform a series of simulation experiments with historical data in the NASDAQ and S&P500 markets between 2006 and 2008. The results show that adding a terrain strategy to a previously successful Memetic Algorithm promoted niching and speciation of the population, which led to a significant improvement in the performance when compared to previous evolutionary methods. We also show that the use of Memetic Algorithms gives the evolved solutions a degree of adaptability to changes in a dynamic market.