中文摘要：

大规模多智能体系统是分布式人工智能领域的一个研究热点，由于多智能体问复杂的交互过程受到诸多不确定性因素的影响，实现系统的宏观优化是关键难点．前期研究发现，某一参数的小范围变化可能引起整个多智能体系统性能的大范围的波动；并且多智能体控制协作算法的研究多数仅与非智能的控制方法进行简单对比，缺乏通用的仿真平台来抽象系统环境并模拟复杂的多智能体行为．为此设计了一种通用的大规模多智能体协同控制平台，该平台能够模拟多智能体系统必须完成的协同控制任务，并且通过变换多智能体系统的相关参数模型及加载不同的智能体控制策略，可以模拟和评测系统各项协作性能指标，为未来设计和提升算法性能提供研究基础．

英文摘要：

Large-scale multi-agent systems have been a popular research field in distributed artificial intelligence. A major challenge exists in overall team performance optimization because there are a large number of uncertainties in agents＇ complex processes of interactions. In a previous work by the authors, it was found that a small change of a parameter may influence the performance significantly. On the other hand, existing researches of large-scale multi agent coordination algorithms only compared the performances of the non-intelligent coordination methods, where a general simulation platform is required that can not only abstract the details of the environment but also simulate the complex coordination activities for each agent, a general platform is designed for large-scale multi-agent team coordination. The platform is able to simulate a variety of tasks that are imperative to agents＇ team coordination. In addition, by varying the setting of giving team coordination parameters as well as loading different coordination strategies, it can simulate and evaluate system performance according to different properties and provide the clue to building optimization algorithms in the future.