为了克服微分进化的局部收敛问题,通过模拟游牧民族的迁徙机制,提出一种迁徙策略,将其与差分进化算法相结合,得到一种迁徙差分进化算法新范式,利用集成技术,发挥各种差分进化算法的优点,提高算法的全局搜索能力。通过生物序列模体识别实验,验证了该算法的有效性。
In order to solve the problem of local convergence in differential evolution, this paper proposes migration strategy by simulating nomadic migration, and gets a novel migration differential evolution model by merging migration strategy. The algorithm with ensemble technique sufficiently exerts the advantages of different differential evolution, and its global search capability is enhanced badly. The algorithm is used to deal with biological sequence motif detection, and experiments show that it is effective