作　　者： ; ; ; ; (陈桂生）;

机构地区： 北京航空航天大学计算机学院软件开发环境国家重点实验室

出　　处： 《系统仿真学报》 2009年第18期5794-5800,共7页

摘　　要： 为了从宏观上更好的描述嵌入式构件系统的结构及其在细粒度层次上隐藏的特征信息,从一个新的视角将大规模嵌入式构件系统抽象为一个由大量的构件及其关系组成的复杂网络,提出嵌入式构件网络的概念,以嵌入式操作系统eCos为载体,证实了该网络具有小世界、无标度、高积聚等重要特性。基于数据场拓扑势的思想,提出拓扑势中心性的评价方法和嵌入式构件网络的节点重要性排序算法,不仅能提供统一的描述框架,而且可以产生更符合网络拓扑特性的排序结果,对嵌入式构件系统的质量度量标准具有深远含义。 In order to characterize the architecture and hidden properties of embedded component system at the fine-grained level, a new viewpoint that the embedded component system with large scale could be studied as a large amount of components connected by the relationships. Thus the Embedded Component Network （ECN） was named. Then taking the eCos for an example, the experiment results confirm that the network exhibits some characteristics of small-world, scale-free and high-cohesion. As a result, a general measure of potential centrality and a new algorithm to identify important components in the ECN based on data field potential were proposed, which contribute to the quality measurement of embedded component system.

关 键 词： 构件复用 复杂网络 度分布 无标度 小世界 嵌入式系统

领　　域： [自然科学总论] [自动化与计算机技术] [自动化与计算机技术]