Fuzzy Flower Pollination Algorithm with Chaos for Global Optimization

Abstract

This paper is concerned with development of a hybrid flower pollination algorithm by combining fuzzy logic and chaos theory. Flower pollination algorithm is one of the recently developed algorithms which mimics the pollination process of the flowers observed in nature. Flower pollination algorithm has two main procedures which are global and local pollination, respectively. These two mechanisms are guided by a switching probability which is a user-defined parameter. Setting a proper value for switching probability is a challenging issue. In this paper, fuzzy inference system is used to control switching probability by considering generation number and population diversity. Furthermore, chaotic numbers are utilized to enhance search capability of the algorithm. All of the proposed modifications are tested on unconstrained function minimization problems and non-parametric statistical tests are used to validate improvements over canonical flower pollination algorithm.KeywordsFlower Pollination AlgorithmFuzzy inference systemChaotic numbersGlobal optimization