A new three-dimensional encoding multiobjective evolutionary algorithm with application to the portfolio optimization problem.

Abstract The existing evolutionary algorithm techniques have limited capabilities in solving large-scale combinatorial problems due to their large search space, making impractical the examination of big real-world instances. In this paper, we address this issue by introducing a new algorithm that incorporates a coding structure specially designed to keep the processing time invariant to the size of the examined test instance, allowing the consideration of large-scale problems for a fraction of time required by other techniques. We test the performance of the proposed algorithm to the optimal allocation of limited resources to a number of competing investment opportunities for optimizing the objectives. We believe that the proposed algorithm can be particularly useful in other contexts too, subject to adaptations relevant to specific problem requirements. Highlights • Existing techniques have limited capabilities in solving large combinatorial problems. • The proposed algorithm keeps the processing time invariant to the problem's size. • It is tested to optimal allocation of limited resources to a number of investments. • It outperforms the other techniques in terms of performance and computational time. • It can be proved very useful in problems with large number of alternative choices. [ABSTRACT FROM AUTHOR]