Showing papers on "Network theory published in 1993"

Centrality in Social Networks

Abstract: In the paper an introduction to main social networks centrality measures is given. A new view on these measures is proposed, based on relational algebra . All described measures are implemented in computer programs CENTRAL and FLOWIND.

Graph & network theory and cellular mobile communications

Abstract: The authors review the application of graph and network theory to mobile communication, especially, to channel assignment problems in cellular systems. Recently, two types of coloring problems on graphs or networks related to the channel assignment problem were proposed. These coloring problems are introduced. Some results on these problems are shown. The relation between cell structure and the Voronoi diagram is considered. >

A new general state-space representation for discrete-time systems

Abstract: A new general state-space representation for causal and anticausal discrete-time systems is introduced. Some of its properties related to systems and network theory concepts are developed. The relation to the descriptor form state-space representation is also explained. Applications are given to show the usefulness of the new representation.

Process Network Theory and Implementation for Technology Assessment in Manufacturing

Abstract: This paper describes a theoretical framework called Process Network Theory and its implementation for evaluating the economic, technical and environmental performance of manufacturing systems. Historically, performance of manufacturing enterprises has been measured largely in monetary terms such as cash, value added and net return on investment. Achieving and maintaining a long-term competitive position in a global economy requires looking beyond monetary performance measures and financial accounting systems. It is necessary to evaluate alternative technologies in the context of their economic performance, technical performance and environmental loading. Technical performance, as defined here is the physical performance of manufacturing systems in terms of physical flows of materials, products and energy. Many companies recognize the crucial role of technical performance measures such as defect and scrap rates, yields, throughput time and resource efficiencies in retaining competitive position in the economy. They are experimenting with processes and procedures that stress quality control (TQC), reduce throughput time (JIT), improve process efficiencies, and others. Process network theory provides the ability to determine the technical and economic performance of manufacturing systems by following a two step procedure. As a first step, the physical part of the manufacturing system is represented as a network of elemental processes whose basic function is to transform the technical state of materials using physical energy and skill-specific human time. The models of the elemental processes are parameterized by a set of feasible technologies which are selected based on constraints and opportunities available for the exchange of materials and energies between the manufacturing system and its economic and natural environment. Linear graph theory is used to obtain the overall model of the manufacturing system as a function of the technologies and the network structure. The model is then used to obtain technical, economic and environmental performance as a function of the technologies involved.

Performance evaluation of intelligent network services

Abstract: The intelligent network (IN) concept provides a new architectural framework which allows rapid introduction of new and complex services. Service completion is carried out over a network of physically distributed elements, connected by a C7 signalling network. The impact of the IN-concept on service and network performance is not obvious. Also, the fact that the IN-concept has not been introduced in the Netherlands yet precludes measuring the performance. Therefore the authors have made a generic model of the IN-concept which is presented in this paper. The model is based on queueing network theory and contains the IN physical elements together with the C7 signalling aspects. By implementing the model in a simulation environment, a tool has been constructed for planning and design purposes.

Performance evaluation of intelligent network services

Abstract: The intelligent network (IN) concept provides a new architectural framework which allows rapid introduction of new and complex services. Service completion is carried out over a network of physically distributed elements, connected by a C7 signalling network. The impact of the IN-concept on service and network performance is not obvious. Also, the fact that the IN-concept has not been introduced in the Netherlands yet precludes measuring the performance. Therefore the authors have made a generic model of the IN-concept which is presented in this paper. The model is based on queueing network theory and contains the IN physical elements together with the C7 signalling aspects. By implementing the model in a simulation environment, a tool has been constructed for planning and design purposes.

Network theory of information engines

Abstract: The author deals with a formulation of a network theory of a system of information engines, where information is transmitted, processed and memorized. A deterministic procedure applied to information by an information engine is defined as a logical work, and two measures, information quantity and information vitality, are introduced. The power of an information engine is defined as the maximum capacity of the logical works performed by the engine, and important properties of total power of the information network, a system of interconnected information engines, are derived. The minimizing problems of power and cost of information networks are discussed. >

Sequential and parallel block-structured shortest-path ray-tracing algorithms

Abstract: Shortest path network algorithms have been applied to geophysical raytracing by Nakanishi and Yamaguchi7 and Moser6 . In the present work two new algorithms are introduced which further exploit the special structure ofnetwork models as applied to geological structures. The algorithms differ in the scanning order of network blocks ina processing queue. An early version of this concept, with computed examples was given'2.Mathematical correctness of both algorithms, "block-greedy" and "block-breadth" , is established. It is then shown that block-greedy limits the updating cycles per block but is inherently sequential whereas block-breadthmay incur back-tracking but allows parallel execution. Shared memory architectures allow block level parallelizationwith guided self-scheduling on a global queue. Message-passing architectures allow direct parallel counterparts tosequential breadth-first executioii. 1. INTRODUCTION Traditionally, geophysists have computed seismic raypaths between given points by shooting methods or bybending methods"4'10. It was not until 1986 that shortest path algorithms from network theory were discovered tohave utility in this context. Dijkstra's method3 was applied to networks of nodes which were imbedded in an arrayof square cells representing discrete blocks of acoustically propagating materials in which the velocity coefficient wasassumed to be constant within each cell, but variable from cell to cell7. An improved, faster, algorithm which utilizedthe heap data structure was proposed6.Networks of this type are more specialized than general networks. Direct connections between nodes are restrictedto nodes within the same cell which are uniformly spaced around its perimeter. Each network edge value is the timerequired for an acoustic ray to travel from one node to the other. The nodes have been points in a plane in applicationsto date but the concept is easily modifiable to points in three space.The importance of the application leads to the consideration of improved shortest path algorithms in networkswhich have a regular structure and in which the edge values (costs) have known constraints. Here, we improve onMoser's algorithm. We use the same network structure but we introduce the concept of processing by cell or blockrather than by individual node. An early version of this concept, with computed examples had been given'2.Since modification for 3-D networks is obvious, in our discussion graph G with node set V and edge set E will beused to represent a 2-D rectangular model, consisting of tiled square blocks such that acoustic velocity is constantwithin each block. In this tiled grid, h + 1 nodes are evenly placed on a single boundary line of each block, one nodeat each corner and h —