The application of water cycle algorithm to portfolio selection

Abstract

Portfolio selection is one of the most vital financial problems in literature. The studied problem is a nonlinear multi-objective problem which has been solved by a variety of heuristic and metaheuristic techniques. In this article, a metaheuristic optimiser, the multiobjective water cycle algorithm (MOWCA), is represented to find efficient frontiers associated with the standard mean-variance (MV) portfolio optimisation model. The inspired concept of WCA is based on the simulation of water cycle process in the nature. Computational results are obtained for analyses of daily data for the period January 2012 to December 2014, including S&P100 in the US, Hang Seng in Hong Kong, FTSE100 in the UK, and DAX100 in Germany. The performance of the MOWCA for solving portfolio optimisation problems has been evaluated in comparison with other multi-objective optimisers including the NSGA-II and multiobjective particle swarm optimisation (MOPSO). Four well-known performance metrics are used to compare the reported optimisers. Statistical optimisation results indicate that the applied MOWCA is an efficient and practical optimiser compared with the other methods for handling portfolio optimisation problems.