Modified Flower Pollination Algorithm for Function Optimization

The flower pollination algorithm (FPA) was firstly proposed in 2012 as one of the population-based metaheuristic optimization search techniques. It is conceptualized by the pollination behavior of flowering plants. In this paper, the new enhanced version of the original FPA named the modified flower pollination algorithm (MoFPA) is proposed to improve its search performance for function optimization. The switching probability of the original FPA used for selection between local and global pollinations is changed from the fixed manner to the random manner according to the pollination behavior of flowering plants in nature. To perform its effectiveness, the proposed MoFPA is tested against ten standard benchmark functions compared with the original FPA. As simulation results, it was found that the proposed MoFPA performs superior search performance for function optimization to the original FPA with higher success rates and faster search time consumed.