【作者】 朱莉；

【导师】 齐欢；

【作者基本信息】 华中科技大学 ， 系统工程， 2006， 硕士

【摘要】 随着经济和社会的发展,群体行为理论在生物、社会、经济以及管理等领域日益显示出其适用性和研究价值,正日益受到各界的重视。在复杂多样的群体行为中,本文选择了基础并且具有代表意义的队列行为,希望能通过对基于局部信息的群体队列行为进行试探性研究,为以后的群体行为研究打下基础。本文利用复杂适应系统的Swarm平台提供的环境构造出队列以及队列中的个体,利用Swarm平台对其中Agent的控制建立队列中个体的调整规则,利用个体运动导致的队列整体行为变化来得出对现实相关领域有用的结论。从课题提出、系统总体设计、详细设计、实现到仿真结果的分析,成功地利用Swarm仿真平台实现了基于局部信息的群体队列行为模拟系统,仿真系统的运行状况和最后结果基本达到了预定的目标。该系统具有以下特征:通过对队列群体中个体运行算法的设计,让个体在局部信息的基础上进行调整,从而使队列整体的状态发生改变,实现了仿真系统的协调推进;采用图形界面显示个体的调整过程,直观清楚地显示了系统整体行为变化的整个过程;个体调整规则经过仔细设计,能得到较好的仿真结果;采用面向对象地语言实现仿真系统具有较强地可靠性和可扩展性。由于本课题只是利用Swarm仿真平台对群体队列行为进行试探性的研究,因此系统仍然有欠完善之处,日后可逐步改善。本文采用面向对象的Java语言在Eclipse3.0的环境下搭建了复杂适应系统的Swarm仿真平台,以实现基于局部信息的群体队列行为系统。更多还原

【Abstract】 With the development of economy and society, group behavior theory has attracted increasing attention due to its applicability and value in the fields of biology, society, economy and management. In the various research fields of group behaviors, we selected the elementary and typical“ranking behavior”as our research field.According to the simulation development theory, an attempted research was conducted on the ranking behavior based on partial information. From the presentation of subject, system design, detailed design, simulation to the results analysis, a simulation system of ranking behavior based on partial information was successfully built up with the Swarm simulation platform of complex adaptive system. The simulation process and final outcome have basically meet expectations.The system has the following features: Individuals adjusted their behavior on the basis of partial information according to the set running algorithm, which resulted in the general adjustment of group behavior, and the coordinated processing of simulation; Group behavior, as well as the individual adjustment, was clearly displayed through the graphical interface; Rules of the individual adjustment were carefully designed so that better simulation results can be obtained; Object-oriented language was adopted so that the simulation system will have better reliability and expansibility. However, since it’s only an attempted research on the ranking behavior, the simulation system still needs improving in later research.To realize the simulation of ranking behavior based on partial information, the object-oriented Java language was adopted to build up the Swarm simulation platform under the environment of Eclipse3.0.更多还原