A probability-based core dandelion guided dandelion algorithm and application to traffic flow prediction

The Dandelion Algorithm (DA) is a recently proposed intelligent optimization algorithm inspired by dandelion sowing. For enhancing its exploitation ability and speeding up its convergence, this work proposes a probability-based core dandelion guided dandelion algorithm (GDA). Specifically, the probability of dandelions being selected is calculated firstly. Then the dandelions need to learn from the core dandelion based on previously calculated probability in the process of sowing. Meanwhile, a greedy selection strategy is applied to GDA. Experimental results show that the proposed algorithm not only outperforms DA and its variants, but also outperforms eight state-of-the-art intelligent optimization algorithms on most functions and three real world problems. In addition, the proposed algorithm is applied to optimize kernel extreme learning machine (KELM) for traffic flow prediction, and the results show that the proposed model has a smaller prediction error than its peers.