Research And Application On Improved Artificial Fish Swarm Algorithm Based On Cloud Model

Artificial fish swarm algorithm is an efficient optimization algorithm. It is simple and easy to operate, robust, and easy to combine with other method so that it was applied to many fields. In addition, it is limited to the generation of artificial fish model. Algorithm only considers the tendency in the evolution of fish and ignores the randomness. So fish evolution and compatibility issues related to the randomness of the tendency has become a research trend. To solve this problem, use the ability of converting the uncertain relationship between qualitative and quantitative and propose an improved algorithm based on cloud model. Clouds learning factor and cloud variability factor are introduced to the algorithm. Improve the ability of artificial fish in the optimization process of active learning. Effectively avoid the fuzziness of the artificial fish swimming behavior in the optimization and improve the performance of the algorithm optimization. In addressing multi-threshold image problem, due to the traditional algorithms can not reasonably lead to the selection of threshold segmentation inaccuracies. To address this uncertainty threshold issue, the improved algorithm is applied to multi-threshold image segmentation problem. Combined with maximum entropy function, adjust fish parameters adaptively. Experiments show that the improved fish image segmentation algorithm ensure the accuracy and stability. Improve the speed of image segmentation at the same time.