基于Attention注意力机制下的鱼群韧性研究  

Research on fish resilience based on attention mechanism

作　　者：杜常金 刘磊[1,2] DU Changjin;LIU Lei(Business School,University of Shanghai for Science and Technology,Shanghai 200093,China;School of Optical-Electrical,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学管理学院,上海200093 [2]上海理工大学光电学院,上海200093

出　　处：《智能计算机与应用》2024年第4期209-214,共6页Intelligent Computer and Applications

摘　　要：生物集群运动可以自组织实现群体涌现行为,但是在外界因素的影响之下,生物集群是否能够保持韧性,单体是否能够根据当前信息重新交互仍然面临巨大的挑战。在攻击体的干扰之下,本文根据红鼻剪刀鱼的运动数据,设计Attention注意力模型。模型考虑到单体鱼与攻击体和周围邻居鱼之间的信息交互,预测下一时刻单体出现的具体位置,说明Attention注意力模型能够使生物的集群韧性与自组织运动保持一致。实验结果表明,所提的Attention注意力模型能够较好的解释生物集群韧性,增强生物集群韧性的鲁棒性和灵活性,为复杂系统解释内外部之间的联系提供了有力的支撑,该方法对生物控制领域的分布式管理也有很好的借鉴作用。Biological cluster movement can self-organize to realize group emergence behavior,but under the influence of external factors,whether biological cluster can maintain resilience and whether monomer can interact according to current information still face great challenges.Under the interference of the attack body,the Attention attention model was designed according to the movement data of the red-nose scissor fish.The model takes into account the information interaction between the monomer fish and the attacking body and the surrounding neighbor fish,and predicts the specific location of the monomer appearance at the next moment,indicating that the Attention attention model can keep the cluster toughness of the organism consistent with the self-organized movement.The experimental results show that the proposed Attention attention model can better explain the resilience of biological clusters,enhance the robustness and flexibility of biological cluster resilience,and provide strong support for complex systems to explain the internal and external links.This method also has a good reference for distributed management in the field of biological control.

关 键 词：复杂系统控制 Attention注意力模型 生物集群韧性 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]