变尺度动态网格法在数据分发管理中的研究

Dynamic Grid-based Method Research in Data Distributed Management

在分析高层体系结构下数据分发管理机制实现的各种静态和动态方法的基础上,指出了仿真过程中时间开销与数据过滤率是一对矛盾的因素,提高数据过滤率必然导致时间开销的增加。一个仿真系统不能只采用一种固定的网格单元尺寸,而要随系统的不同动态改变网格的单元尺寸,以达到在最短时间消耗情况下得到最高的数据过滤率。由此提出了一种变尺度动态网格法,在深入分析网格单元尺寸与更新时间、数据过滤率、接收时间及排队时间的关系后,给出了该算法的数据方程及实现过程;最后在一个应用实例中对该算法进行了验证,说明了该算法的有效性。

This paper points out that time consuming and data filtering are a pair of element contradiction each other in process of simulation, analyzing static and dynamic methods of data distributed management mechanism based on high level architecture, the higher rate of data filter is improved, the more time will be consumed, vice versa. A distributed simulation system should have a dynamic grid-based method in data distributed management, but not a static grid, thus the system can get a high rate of data filter on a low consume of time. So the article presents a method of dynamic grid-based method and the construction equation of the method and the process of developing it based on further analyzing the relationship between the update time and data filer, receiving time and queue time, and verifies the availability of the algorithm using an example.