On the performances of the flower pollination algorithm - Qualitative and quantitative analyses

The flower pollination algorithm (FPA) is a recently proposed metaheuristic inspired by the natural phenomenon of flower pollination. Since its invention, this star-rising metaheuristic has gained a big interest in the community of metaheuristic optimisation. So, many works based on the FPA have already been proposed. However, these works have not given any deep analysis of the performances of the basic algorithm, neither of the variants already proposed. This makes it difficult to decide on the applicability of this new metaheuristic in real-world applications. This paper qualitatively and quantitatively analyses this metaheuristic. The qualitative analysis studies the basic variant of the FPA and some extensions of it. For quantitative analysis, the FPA is statistically examined through using it to solve the CEC 2013 benchmarks for real-parameter continuous optimisation, then by applying it on some of the CEC 2011 benchmarks for real-world optimisation problems. In addition, some extensions of the FPA, based on opposition-based learning and the modification of the movement equation in the global-pollination operator, are presented and also analysed in this paper. On the whole, the basic FPA has been found to offer less than average performances when compared to state-of-the-art algorithms, and the best of the proposed extensions has reached average results. (C) 2015 Elsevier B.V. All rights reserved.