多个体集群运动中信息共享引致对称破缺的模型与实验

Collective motion, as a kind of collective behavior, is a typical complex system. The study on information use in collective behavior is just at the beginning. The preliminary results in our ant colony panic experiment showed the symmetry breaking vs the ant number is far from that given by the frequently adopted Vicsek model. The group direction change in some ecological groups can not be explained either by Vicsek model. We find that it is information sharing that resulted in the ant colony symmetry breaking, and for the latter, it is symmetry breaking that leads to the change of group directions. Only by studying information sharing and symmetry breaking, can we solve these problems. It is a more essential scientific problem to study the symmetry breaking induced by information sharing that leads to alignment and consensus in collective behavior. This proposed project will study the symmetry breaking induced by information sharing. Firstly, this project sets up a model of pheromone that leads to symmetry breaking to explain results from the ant colony panic experiment. Secondly, the symmetry breaking is used to describe the group direction change and the fish school experiment will be committed to verify the model. Thirdly, information and symmetry breaking will be used to increase the survival time of the escapees in the chase and escape model. Finally, we will also study the least number of controllers and control energy of Multi-agent systems to arrive at any state resulted from the symmetry breaking by combining researches on the inferring complex networks from time series and the dynamic network control.

集群运动是基于多个体系统集体行为的一种，是复杂系统研究的典型问题之一，而信息在集体行为中功能的研究刚刚开始。我们在蚁群实验中得到蚁群对称性破缺对蚁群密度的变化曲线与已有Vicsek模型结果相去甚远；一些生物群体快速地改变群体运动方向的现象用现有的Vicsek模型也无法解释。我们发现前一实验恰好是信息共享造成了蚁群对称破缺，后一现象也恰是对称性破缺造成了群体方向性的改变。只有研究信息共享与对称性破缺，才有可能对这些问题给出正确答案。信息共享引致的对称破缺造成了集群运动中的对齐与一致性，研究对称性破缺是更基本的科学问题。本项目将研究集体行为中信息共享造成的群体行为对称破缺，包括蚁群对称破缺的信息素模型和实验验证，群体运动方向改变的对称破缺模型与鱼群实验验证，追逃问题的信息和对称破缺利用。并将从数据推定复杂网络结构与动态网络控制问题结合，研究控制多主体系统，达到由对称破缺导致的任意态，所需的最少个体集合和最小能量。

本项目针对多个体集群运动中信息共享引致对称破缺的模型与实验进行研究。项目研究内容包括：蚁群恐慌逃生的对称性破缺实验与模型研究，恐慌情况下人群疏散实验与模型，群体追逃问题中的对称破缺，鱼群u-turn实验与模型，集群行为模式识别与轨迹追踪，以及机器人群体实验（swarm robotics）多主体系统控制；开发了适用于群体行为研究的视频模式识别、轨迹追踪方法；配合实验的进行，进行了机制发掘、模型建立。模型建立同时指导了实验的进行。.通过实验的方法进行复杂系统研究，是艰难的，也是现在国内较少研究组采取的研究路线。通过实验研究也是独特的、有前途并可能取得实质进展、取得原创性成果的研究方法。.科研进展基本达到项目目标，实质国际交流与合作广泛，并形成国际影响，人才培养取得合理进展。.