Performance profile assessment of electromagnetism-like algorithms for global optimization

Abstract

This paper introduces a modification on the movement force vector of the Birbil and Fang's electromagnetism-like algorithm for solving global optimization problems with bounded variables. The proposed movement vector combines the total force exerted on each point of the population, at the current iteration, with the rate of change in the force vector of a previous iteration. Several widely used benchmark problems were solved to compare the proposed modification with the original algorithm. A performance profile assessment shows the efficiency and robustness of the proposed modification.