随着信息科学技术的迅速发展，离散事件系统仿真技术在各个领域有了更深入的研究。在基于离散事件的大规模复杂系统的仿真研究中，由于系统计算复杂，仿真粒度越来越细化，使得系统对计算资源的需求日益增加。缩短大规模复杂仿真系统运行的时间，满足系统的时效性需求，是一个重要的研究方向。

首先，本文调研了描述离散事件仿真系统的形式化的方法：离散事件系统规范(Discrete Event System Specifications, DEVS)，并基于DEVS对大规模复杂仿真系统进行了逻辑建模。

其次，在逻辑建模的基础上，本文分析了大规模复杂仿真系统的特征，并结合Hadoop的MapReduce模型，提出了基于离散事件仿真系统的分布式计算框架，对系统进行了软件建模。

       最后，本文给出了分布式计算框架的设计和实现，包括系统的整体架构和数据流程，以及关键模块的设计与实现，并详尽地描述了整个框架的设计原理和实现细节。

With the rapid development of information science and technology, discrete event system simulation technology has a more in-depth research in various fields. In large and complex systems based on discrete event simulation study, given the computational complexity of the system, coupled with that the the simulation granularity is more and more refined, the demand for more computing resources is increasing. Shortening the time of running the large-scale complex simulation systems to meet the timeliness requirements of the system, is an important research direction.

Firstly, this paper surveys the formal method of the description of discrete event simulation system: Discrete Event System Specification (DEVS) and builds a logical model of the complex simulation system using DEVS.

Secondly, on the basis of logic modeling, this paper analyzes the characteristics of the large-scale complex simulation system, combined with the MapReduce model in Hadoop, and proposes a distributed computing framework based on discrete event simulation system, and then builds the software model of the system.

Finally, this paper provides the design and realization of the distributed computing framework, including overall system architecture and data flow, as well as key modules design and implementation of the framework and detailed description of the design principles and implementation.