Portfolio optimization using novel co-variance guided Artificial Bee Colony algorithm

Although the use of evolutionary algorithms and fuzzy logic for portfolio optimization is an established research area, this field remains fascinating because of its important financial aspects. The field is brisk and it trances as there always remain research issues which are yet to explore. The problem of portfolio optimization comprises of finding an optimal distribution of funds among various available securities so as to maximize the return and minimize the risk. Artificial Bee Colony (ABC) is one of the effectual and widely used optimization technique based on swarm intelligence. Mixing co-variance principles with ABC algorithm assists in quick convergence with more precision. This paper presents a novel co-variance guided Artificial Bee Colony algorithm for portfolio optimization. As portfolio optimization consists of simultaneous optimization of multiple conflicting objectives, this algorithm is named as Multi-objective Co-variance based ABC (M-CABC). The efficacy of the proposed algorithm is tested on benchmark problems of portfolio optimization from the OR-library. The results validate the adept performance of the proposed algorithm in finding various optimal trade-off solutions simultaneously handling realistic constraints. The article concludes with exhaustive post-result analysis and observatory remarks to bring out some of the crucial properties of optimal portfolios.