A novel hybrid nature-inspired scheme for solving a financial optimization problem

Abstract

Hybrid intelligent approaches have proven their potential in demanding problem settings. The financial domain provides some challenging problem set-ups, mostly because of non-linearity conditions and conflicting objectives and binding restrictions. In this study, a novel hybrid algorithm, which stems from Nature-Inspired Intelligence, is applied in a specific portfolio optimization problem. The proposed algorithm comprises of an Ant Colony Optimization Algorithm (ACO) for detecting optimal combination of assets and a Gravitational Search Algorithm (GSA), for optimal capital allocation in the portfolio. Results from the proposed hybrid scheme are compared to previous findings, in the same optimization problem and dataset, from another hybrid NII algorithm, namely ACO Algorithm with Firefly Algorithm (FA). Experimental findings indicate that the proposed hybrid scheme yields a promising distribution of fitness values from independent simulation runs. What is more, in terms of best solution found, the proposed hybrid scheme yielded a solution that is 7.2% worst than the benchmark approach’s one. However, in terms of execution time, the proposed algorithm was faster. Taking into consideration both the above aspects, the difference of the two hybrid algorithms, in terms of best solution, can be characterized as insignificant. The main aim of the paper is to highlight the advantages of the proposed hybrid scheme, as well as the great potential of Nature-Inspired Intelligent algorithms for the financial portfolio optimization problem.