Survey of oppositional algorithms

Evolutionary algorithms (EA) are gaining popularity due to their success in solving optimization problems. New methods are being introduced continuously in the literature. One such technique, biogeography-based optimization (BBO) is proving itself as an effective algorithm for real-world problems with large number of variables. Previous work shows that augmenting BBO and other EA with opposition-based learning increases their performance, specially for high-dimensional problems. Five different oppositional algorithms have already been separately introduced in the literature. This paper combines BBO with these five oppositional methods and tests their performance on 21 benchmark problems. Furthermore, a new oppositional algorithm, fitness-ranking-based central opposition (FCB), is introduced for the first time. FCB was able to solve two of the problems that other EA could not. FCB is also determined to be the oppositional algorithm with the highest success rate.