Optimising a targeted fund of strategies using genetic algorithms

Abstract

This paper examines the use of Genetic Algorithm in order to perform the task of continuously rebalancing a portfolio targeting specific risk and return characteristics. The portfolio is comprised of a number of arbitrarily performing trading strategies, plus a risk-free strategy in order to rebalance in a similar method to the traditional CAPM method of rebalancing portfolios. A format is proposed for designing a fitness function appropriate to the task, and evaluated through the final results. Results of targeting both risk and return are investigated and compared, as well as optimising the non-targeted variable in order to create efficient portfolios. The findings show that a Genetic Algorithm is indeed a viable tool for optimising a targeted portfolio, using the proposed fitness function.