Graph Structures for Discrete Event Simulation

仿真可能是现代科学和工程研究中最重要的计算应用。仿真模型有两大类：连续仿真和离散事件仿真。连续模拟是这两种方法中比较流行的一种，很大程度上是由于它们以微分或差分方程的形式显式的数学结构。虽然离散事件模拟在通用性和计算效率方面有一些优势，但它们还没有明确的数学表示，接近微分方程的通用性和实用性。离散事件模拟缺乏数学结构，这在一定程度上说明了这些模型具有明显的即时性和复杂性。离散事件模拟通常很难正确编程。最终的研究目标是发展一个数学基础，将在离散事件模拟模型的开发和实验中具有实际应用。

Simulation is potentially the most important computing application in modern scientific and engineering research. There are two broad classifications of simulation models: continuous simulations and discrete event simulations. Continuous simulations are the more popular of the two approaches largely due their explicit mathematical structure in the form of differential or difference equations. While discrete event simulations have some advantages in terms of generality and computational efficiency, they do not yet have an explicit mathematical representation with anywhere near the generality and utility of differential equations. The lack of mathematical structure for discrete event simulations accounts in part for the apparent ad- hoc and complex nature of these models. Discrete event simulations are typically very difficult to program correctly. The ultimate research goal is the development of a mathematical base that will have practical applications in developing and experimenting with discrete event simulation models.