Chaos and intensification enhanced flower pollination algorithm to solve mechanical design and unconstrained function optimization problems

Abstract

Nature-inspired computation has enjoyed a visible position among the soft computational techniques. Flower Pollination Algorithm (FPA), which is known as one of the outstanding algorithms in this domain, has been shown to be promising in numerous publications. FPA is comprised of two main phases, which are referred to as abiotic and biotic pollination that correspond to local and global search, respectively. It makes use of a user-supplied parameter to switch between them. This parameter is referred to as the switching probability. Thus, it can be put forward that switching probability defines the search characteristic of FPA. The present work introduces several FPA modifications that adopt chaotic maps. Moreover, the developed modifications are further enhanced by using intensifying step sizing procedure that allows a more intensified search towards the end of search. With the help of the introduced chaos in switching probability and incrementally intensifying search, developed FPA modifications are expected to find the hard-to-detect promising regions. Next, such capabilities of chaotic maps are utilized in various building blocks of FPA. Performances of all developed FPA modifications are analysed on the well-known unconstrained real-valued function minimization and mechanical design problems. Finally, appropriate non-parametric statistical analysis is carried out to observe possible statistically significant improvements over the standard FPA. As shown by the experimental study, obtained results induce success of the developed procedures, which clearly add to the capability of the canonical FPA.