基于Agent的建模与仿真和离散事件仿真的集成应用研究

基于Agent的建模与仿真(ABMS)和离散事件仿真(DES)同属先进的建模仿真模式,都具备一定处理复杂系统问题的能力。ABMS继承面向对象思想,擅长从微观角度着手分析复杂系统;DES沿用传统的面向过程思想,能对复杂系统整体的宏观行为进行分析。二者在功能上具有互补性。在对两种模式深入分析的基础上,本文提出将二者集成应用的思想,扩展了传统DES事件表的结构,提出将ABMS和DES集成应用的完整框架,并实现了一个支持此框架实现的软件平台——absimlib,充分发挥了两种仿真模式的优势,弥补了单独使用其中任何一种的不足,增强了对复杂系统的建模与仿真能力,且结构简单、易于实现和向分布式环境下扩展。最后以一个防空作战的例子证明此方法的有效性。

Study on Integrated Usage of Agent-based Modeling and Simulation and Discrete Event Simulation

Agent-based Modeling and Simulation (ABMS) and Discrete Event Simulation (DES) are all advanced modeling and simulation paradigms, and they have abilities for analysis of the problems about complex systems. ABMS follows the object-oriented ideas, so it can work well in the micro aspect of systems; and DES adopts progress-oriented ideas, so it can work well in the macro aspect. Therefore, they can amend each other on the functions. By analyzing them in details, the idea was proposed to integrate them for using, which extends the structure of event list in DES, brings out a full architecture for integrated usage of ABMS and DES, and realizes a software platform, absimlib, to support the realization of the architecture. All the efforts bring into playing the advantages of the two simulation paradigms, avoid the shortages incurred by using one of them alone, and increase the abilities for studying on complex systems. Furthermore, the architecture is simple, and easy to realize and extend to distributed environments. At last, an air-defense system was given out to show the effectiveness of the idea as an example.