A graph-based expert system approach to geometric feature recognition

Geometric feature recognition is a crucial task in the development of concurrent engineering software. This paper presents a new methodology for geometric feature recognition which combines the advantages of face-edge adjacency graphs and expert systems. The methodology uses several new concepts such as enhanced winged edge data structure (EWEDS) and multi-attributed adjacency graphs (MAAG). The object model is presented as a set of facts. The rules for the recognition of each feature are derived from the corresponding feature-MAAG. This simplifies the process of writing the rules while enabling the inclusion of new features into the rule base as they are encountered.     

连通图的邻接矩阵判定法

利用图邻接矩阵的定义及性质,得到了利用邻接矩阵来判定无向图为连通图及非连通图的判定定理,并通过例题说明了利用该定理来判定连通图的具体解题过程.     

基于预测信息的时间依赖网络路径规划算法

分析最小时间规划算法,针对算法在应用上存在的不足,在研究SPFA静态规划算法的基础上,结合该规划算法的优点,提出一种改进的最小时间路径规划算法。实例验证结果表明,该算法能够提供实时、高效、预测性强的规划路径,在城市交通中较好地满足用户的需求。     

一种基于最少边数的最短路径算法

确定两点之间最短路径,通常要求该路径满足两点之间的权之和最小。为此采用层次遍历图的思想,设计了一种新的结构存放路径选择信息,找到一种确定这种最短路径的算法,并给出了算法描述以及实例。     

基于谱约束的随机化社会网络多点扰动方法

现有基于谱约束的随机化社会网络扰动方法只采用4个点的扰动,扰动后社会网络的隐私保护程度不强。为此,基于邻接矩阵及无符号拉普拉斯矩阵,提出一种多点扰动方法,在随机化过程中将社会网络的谱半径控制在一定约束范围内,能在保证扰动后社会网络可用性的同时提高其隐私保护程度。实验结果表明,该方法可以更好地保护社会网络结构。     

The adjacency graph of a real algebraic surface

The paper deals with the question of recognizing the mutual positions of the connected components of a non-singular real projective surface S in the real projective 3-space. We present an algorithm that answers this question through the computation of the adjacency graph of the surface; it also allows to decide whether each connected component is contractible or not. The algorithm, combined with a previous one returning as an output the topology of the surface, computes a set of data invariant up to ambient-homeomorphism which, though not sufficient to determine the pair , give information about the nature of the surface as an embedded object. This research was partially performed with the contribution of M.I.U.R. and of Eurocontract HPRN-CT-2001-00271     

基于邻接矩阵的低压配电网故障区段定位方法

低压配电网具有线路区段短、分支众多且拓扑结构复杂多变的特点,分布式电源的并网进一步增加了其故障定位的难度。结合图论分析方法和人工智能算法,提出了基于拓扑邻接矩阵的低压配电网故障区段定位方法。首先生成与配电网故障监测单元配置相对应的邻接矩阵,其次建立故障电流在故障监测单元拓扑中的连通性等效模型,根据故障电流从电源到估计故障区段的最短路径计算故障监测信息。以此为基础建立了考虑多重故障的低压配电网故障定位模型并利用粒子群优化算法进行模型求解。仿真算例验证了所提方法能够有效提高故障定位的效率和准确度。     

基于拓扑片的智能分布式馈线拓扑分析

馈线拓扑分析是实现馈线自动化的基础,智能分布式馈线自动化(Distributed Intelligent Feeder Automation,DIFA)的馈线拓扑分析由配电终端自主实现面临着巨大的挑战。首先研究了FA应用的馈线拓扑模型,把电源、线路、母线和负荷等电气元件抽象为节点,以开关为边-电气元件为节点的多叉树表示的馈线拓扑作为DIFA馈线拓扑模型。提出了基于拓扑片的馈线拓扑分散配置方法,把每个配电终端覆盖的区域定义为拓扑片,每个配电终端配置存储拓扑片信息。把拓扑片的邻接关系用配电终端的邻接关系来定义,并映射成开关邻接关系。在拓扑片的配置中显性配置开关邻接关系,隐性包含线路节点、母线节点和负荷节点。最后提出了拓扑片拼接的馈线拓扑静态分析方法。所提出的基于拓扑片的馈线拓扑分析方法适用于DIFA,由配电终端自主实现,具有分析速度快和配置简单的优势。     

Constructing an optimal facility layout to maximize adjacency as a function of common boundary length

An effective facility layout implies that departments with high flow are laid adjacent. However, in the case of a very narrow boundary length between the neighbouring departments, the adjacency would actually be useless. In traditional layout design methods, a score is generally assigned independent of the department’s boundary length. This may result in a layout design with a restricted material flow. This article proposes a new concept of adjacency in which the department pairs are laid adjacent with a wider path. To apply this concept, a shop with unequal rectangular departments is contemplated and a mathematical programming model with the objective of maximizing the sum of the adjacency degrees is proposed. A computational experiment is conducted to demonstrate the efficiency of the layout design. It is demonstrated that the new concept provides a more efficient and a more realistic layout design.