改进的基于排序的DDM算法  

An Improved Sort-based DDM Algorithm

作　　者：张霞[1] 延耀威 ZHANG Xia;YAN Yaowei(School of Computer and Information Technology,Shanxi University,Taiyuan 030006,China)

机构地区：[1]山西大学计算机与信息技术学院,山西太原030006

出　　处：《山西大学学报（自然科学版）》2019年第1期120-124,共5页Journal of Shanxi University(Natural Science Edition)

基　　金：山西省自然科学基金(201701D121051)

摘　　要：HLA(High Level Architecture,高层体系架构)是目前大规模仿真中广泛应用的仿真框架。随着仿真规模的逐渐扩大,各个仿真实体之间的交互信息越来越多,如果不对交互信息进行数据过滤,大量不必要的信息交互将会导致网络资源匮乏,仿真延迟增加,仿真系统的可扩缩性降低,所以仿真节点间的数据分发管理越来越重要。本文针对DDM(Date Distribution Management,数据分发管理)算法中基于排序的算法进行了改进,提出将订购区域与发布区域分开存储,然后将订购区域的坐标值进行排序,将发布区域中的坐标值依次与订购区域中的坐标值相匹配,减少了原算法中大量的判断过程,实验结果表明改进算法的仿真效率提高比例稳定在9%左右,较原算法有显著提高。HLA is a widely used simulation framework in large-scale simulations.With the gradual expansion of the scale of simulation of the interaction,more and more information appear between various simulation entities.If do not manage the information of data,a lot of unnecessary information interaction will result in the lack of network resources,the increase of simulation latency,and the reduction of the scalability of the simulation system,therefore,the simulation nodes’ data distribution management is becoming more and more important.This paper improves the sort-based algorithm in DDM algorithms.It is proposed that the subscription areas and the publish areas should be stored separately,and the coordinate values of the subscription area are sorted.The coordinate values in the publish areas are matched with the values in the subscription areas in order to reduce a lot of judgement processes in the original algorithm.The experimental results show that the simulation efficiency of the improved algorithm is improved by about 9%,which is significantly higer than that of the original algorithm.

关 键 词：HLA 数据分发管理 基于排序 仿真效率 

分 类 号：TP15[自动化与计算机技术—控制理论与控制工程]