Showing papers on "Layer (object-oriented design) published in 1993"

System for providing a uniform external interface for an object oriented computing system

Abstract: Objects import data from, and export data to, a plurality of external systems in an object oriented computing system via a uniform external interface. The uniform interface includes external interface layers and internal interface layers. Each external interface layer converts data from the associated external system into an identifying token and data value. The external layer also converts an identifying token and data value into data which is formatted for the associated external system. The internal interface layer accepts an identifying token and data value from the external interface layer and converts the data value to a data attribute associated with an object. The internal interface layer also converts a data attribute into an identifying token and data value and provides the token and data value to the external interface layer. Thus, objects need only communicate with the internal interface layer in order to import and export data. The external interface layer includes customized conversion routines for converting a token and data value to data which is useable by an external system. Data validation and storage takes place via the internal interface layer. Accordingly, replication of code is minimized and the complexity of the object oriented system is reduced.

Layer assignment for rubber band routing

Abstract: The flexibility of the rubber-band wire model is very promising for routing and optimizing today''s complex VLSI and MCM. In this paper we described a practical layer assignment algorithm used with a rubber band router. The algorithm uses a steepest descent approach with an heuristic function which estimate the wiring length of the generated assignment. We formulate the cost function and present an efficient way to compute it. The use of the cost function enable to control the final layout and to achieve balance between wire length and number of vias. The algorithm can be use as well for cost-driven one an a half layer designs in which one wiring layer has only short "jumpers" and is embedded in the ground layer.

Method and system for storing and accessing data in a compound document using object linking

Abstract: Through a multi-layer interface, an application program interfaces with a compound document storage system to manipulate compound documents. A first layer provides methods which an application program uses to access a compound document using the functions of a second layer. The second layer maps requests to store data in the compound document to a storage format using the functions of the third layer. The third layer maps requests to write to a file to an arbitrary storage medium.

Nuclear Layer Serialization in Lewo

Abstract: Lewo is spoken on Epi Island in Vanuatu. In Lewo, nuclear layer serialization is very widespread. However, the process of incorporation has progressed to the extent that most of the serialized verbs have become morphologically indistinguishable from a class of verbal suffixes. In 1982, Crowley described for Paamese a class of postverbal adjunct modifier particles. His 1987 reanalysis presented the same form as nuclear layer serialized verbs. He suggests that similar phenomena exist in other languages in the area.

Object-oriented viewing framework having view grouping

Abstract: A view system provides an extensible mechanism for associating a logical set of windows and manipulating them as a unit. For example, operations can be applied across address spaces to all the members of the group. A group is constructed by inserting a reference to each view in the group in a layer object. The layer object, in turn, can be inserted into a data hierarchy structure in a hierarchy object. The data hierarchy structure defines front to back display levels on a display and defines which windows overlap. Since all the members of the group are in the same layer object, they move to different levels as a group. Polymorphism and extensibility are provided via the object-oriented architecture of the operating system.

Topological querying of multiple map layers.

Abstract: This paper first recaptures why multiple map layers are required in Geographic Information Systems. The two main motivations are: flexibility in data modeling, and efficient processing of data. In order to make the map layer discussions clearer, we introduce two different types of map layers: a structure layer, and a thematic layer. Though the concept of a structure layer is defined in a general sense, to illustrate its practicability the organization of data in a structure layer is initially represented according to the formal data structure for single-valued vector maps as proposed by Molenaar. In order to develop a data model for a multi-layered system, the concept of structure layers, specified for the fds, is extended for multi-valued vector maps. It turns out that the data can be modeled in various ways. After that the topic of topological querying of multiple map layers is introduced with a few examples. Map overlay plays a central role in this process. But map overlay is a computationally expensive operation, and therefore several alternative optimization techniques are described for answering the queries efficiently. An important goal of the described multiple map layer query language is that it is a realistic approach. That is, the resulting implementation can be used in an interactive environment with real data sets: with at least several megabytes of geographic data. This is reflected by the case study presented in this paper.

Genetic Search for Optimal Representations in Neural Networks

Abstract: An approach to learning in feed-forward neural networks is put forward that combines gradual synaptic modification at the output layer with genetic adaptation in the lower layer(s). In this “GA-delta” technique, the alleles are linear threshold units (a set of weights and a threshold); a chromosome is a collection of such units, and hence defines a mapping from the input layer to a hidden layer. The fitness is evaluated by measuring the error after a small number of delta rule iterations on the hidden-output weights. Genetic operators are defined on these chromosomes to facilitate search for a mapping that renders the task solvable by a single layer of weights. The performance of GA-delta is presented on several tasks, and the effects of the various operators are analyzed.

Issues in scalable library design for massively parallel computers

Abstract: The authors examine some critical issues raised in the design of libraries for MPCs, such as scalability, portability, recompilation, and flexibility. They advocate a layered structure of library design, comprising a high-level language layer, a machine-independent node layer, a machine-dependent node layer, and an object code layer for different demands and requirements. They discuss the impact of various data decomposition strategies on program performance and the computation and communication analysis techniques used at different layers. They also propose the concept of the range of scalability as a metric for selecting the most appropriate implementation. A linear system solver based on the Gaussian elimination method is used as an example to illustrate various design alternatives.

Optical pattern recognition system based on a winner-take-all model of a neural network

Abstract: A three-layer neural network implemented by optical and electronic techniques for pattern recognition is described. The principle, the architecture, and the experimental results of the hardware system for pattern recognition are presented. A multiplex hologram is used as the interconnections between neurons of the input layer and the middle layer. The winner-take-all (WTA) operation of the intermediate layer is performed by an electronic circuit. The interconnection between neurons of the middle layer and the output layer is simply implemented with an optical transparency. By changing the patterns on the transparency, this system can be used for both auto- or heteroassociative memories. In addition to noise and partial hindrance tolerances, it also has illuminance invariance because of the WTA layer of the system.

Semiconductor memory and fabrication thereof

Abstract: PURPOSE:To realize a nonvolatile semiconductor memory for storing information based on the residual polarization in which high speed operation is allowed by suppressing the wiring delay of word line without causing deterioration in the characteristics of ferroelectric CONSTITUTION:A word line 8 and a data line 12 are provided additionally under a storage capacitor part 22 and a new wiring layer 16 is also provided A hierarchical word line is constituted using the new wiring layer as a word bus thus decreasing the word line delay Since a ferroelectric film is formed after formation of the word bus, deterioration of characteristics does not take place

Learning method/device for neural network

Abstract: PURPOSE:To learn a neural network with high reliability, high efficiency and at a high speed in regard of plural categories. CONSTITUTION:When a specific category to which a subject belongs is identified with use of a three-layer perceptron, this perceptron is first learn in accordance with the category to be identified. At the same time, the (M-1) types of hyper- planes which separate M pieces of categories from each other are formed based on an N-dimensional vector serving as the sample data xi(j) of each category. Then the (M-l) types of parameters forming those hyper-planes are assigned to the (M-1) pieces of units corresponding to a medium layer P2, and the learning is carried out to the layer P2 from an input layer P1. In other words, the coupling weight wn(i) is learnt between the units of both layers P1 and P2 together with the bias value bn of each unit of the layer P2.

Design of multicast multilayered lightwave networks

Abstract: A new architecture for lightwave networks is presented. It takes the form of a multicast multilayered lightwave network. In this architecture the underlying optical layer and the electronic layers work hand in hand to provide both high capacity (in the physical layer) and high connectivity( in the electronic layer). The design and the implementation technique for the multicast multilayered network is studied. The physical layer is a multicast linear lightwave network. The topology of each logical layer is modeled as a hypergraph which is then embedded on the physical layer. These networks can be designed to serve a large number of geographically dispersed users and can meet diverse requirements in terms of service and traffic characteristics. >

Design of embedded hard real-time applications with predictable behaviour

Abstract: In contrast to most of the other studies which only consider selected topic(s) and typically deal with issues of higher levels of design, the research project outlined uses a holistic approach to real-time application design issues. It is set up in layers starting with a predictably behaving computer hardware architecture, followed by an appropriate programming language and, finally, a tool for automated application development. Each of these layers provides a sound basis for the next. For the hardware layer, an asymmetrical parallel architecture is proposed and guidelines for its implementation are given. For the software layer, programming issues are dealt with. A real-time programming language concept is briefly described. Inside the compiler, an analyser is integrated, combining conventional and novel approaches for execution time estimation. Finally, the desired properties of a computer aided hard real-time application design and evaluation tool are enumerated. >

Data control method

Abstract: PURPOSE:To easily decide the state of layer structure from an optional detailed record by providing a layer start record and the detailed record with information on the layer structure. CONSTITUTION:This method is provided with a 1st means which represents the layer by outputting the layer start record forming the layer structure to a work file and a 2nd means which stores the state of elements forming the layer structure in the detailed record as information and outputs it. For example, when an inputted layer state is compared with a last layer state to judge that a new layer is inputted, setting and output processes in steps 105-108 are performed, but when not, setting and output processes in steps 11 and 112 are performed. In a step 106, the detailed record is outputted (updated) and in steps 108 and 109, the layer start record is outputted. Thus, the information on the layer structure is provided to the layer start record and detailed record. Consequently, the state of the layer state can easily be decided from the optional detailed record.

A hierarchical associative memory consisting of multi-layer associative modules

Abstract: The author proposes a hierarchical associative model consisting of multilayer neural associative memory modules which do not have any intra-layer connection and which serve as memories for micro-level retrievals. The constituent memory modules are connected to each other by having their layer in common. The hierarchical architecture decreases the total number of connections as compared with the previously proposed models. The hierarchical model has been applied to a character-word recognition problem and the interaction between the constituent modules has been studied. The success rate of the recognition has been found to be controlled by the model's parameters.

An AI-based design of a driver information unit

Abstract: An artificial intelligence-based approach to the design of a driver information unit (DUI) is described. The process layer used for direct control of effectors and output devices is realized as a computer system equipped with realtime operating system and appropriate software for handling the continuous flow of information from and to the environment. The discreet response layer (DRL) is used for processing symbolic, discrete information originating in the process layer. A symbolic characterization of environment condition is created in the process layer, and every change of this characterization causes a message to be sent to the DRL. The DRL is realized as a rule-based system. The reasoning layer is used for actual reasoning about the environment. The discreet response layer is not able to make decisions other than based on predefined set of behavior rules. When a need for more complex reasoning arises, it has to be realized by the reasoning layer equipped with more sophisticated algorithms. The result is sent back to the DRL for further processing.

Method for obtaining inverse solution by learning of neural network

Abstract: PURPOSE:To provide the method which obtains the inverse solution by learning of a neural network without new development of a system of an inverse model or the like for acquisition of the inverse solution. CONSTITUTION:An auxiliary input layer 111 is provided in the preceding stage of an input layer of a neural network 110, and the input layer 112 and the auxiliary input layer 111 are fixedly coupled, and the neural network 110 is learnt by a learning pattern 120; and after the end of learning, fixed coupling between the input layer 112 and the auxiliary input layer 111 is released, and layers from the input layer 112 to an output layer 114 of the neural network 110 are fixedly coupled, and learning is performed with such learning patterns 120 and 160 that an output pattern 150 has a desired value.

Information processor using neural network

Abstract: PURPOSE:To provide an information processor using a hierarchical neural network which can intuitively analyze data expression in the neural network and shorten processing time necessary for the learning of the neural network. CONSTITUTION:The hierarchical neural network consisting of an input layer 1, an intermediate layer 2 connected to the layer 1 and output layer 3 connected to the layer 2, allowing each of the layers 1 to 3 to have at least one unit or more and mutually connecting respective units among respective layers 1 to 3 is constituted so as to use an optional differentiative function as an I/O function expressing the I/O characteristics of each unit in the layers 2, 3.

Communication data ciphering system between end systems

Abstract: PURPOSE:To ensure the security of the entire network system by applying ciphering processing to a user data only. CONSTITUTION:A data received from a section layer (S layer) consists of an S layer header and an S layer user data. A transport layer (T layer) ciphering sub-layer 8 ciphers the data as a T layer user data 11 of a transport protocol data unit (TPDU) and adds a ciphering header 19 and a T layer header 18 to the data 11 and gives the result to a subordinate layer. A T layer ciphering sub-layer 8 of an opposite end system decodes the user data 11 of the received TPDU fro the subordinate layer 12 and gives the result to an S layer 9 being a host layer.

Applying formal methods to software reuse

Abstract: This dissertation presents an approach, based on formal methods, to the specification, classification, retrieval, and modification of reusable software components. From a set of reusable components that are described by formal specifications, a two-tiered hierarchy of software components is constructed. The hierarchical structure provides a means for representing, storing, browsing, retrieving, and modifying the reusable components; furthermore, the formal specifications provide a means for verifying that a given software component correctly satisfies the current problem. The lower-level hierarchy facilitates the application of logical reasoning techniques for a fine-grained, exact determination of reusable candidates. The higher-level hierarchy provides a coarse-grained determination of reusable candidates. Based upon the framework of the two-tiered component hierarchy, a set of candidate components, which are more general than or analogous to the query specification, can be retrieved from the hierarchy. Two methods are proposed for the modification of candidate components in order to satisfy the query specification. One is to modify the component that is more general than the query specification; the other is to modify a component based on analogy. The graphics-based implementation of the reuse framework is described.

Monitoring system with logical reasoning for mobile robot guidance

Abstract: In this paper we propose a monitoring system for the observation and navigation of mobile vehicles in factory environments. After a short overview about the complete flexible transport system, the monitoring system, consisting of a global sensor system with an integrated logical reasoning control module, is described in more detail. The aim of the global monitoring is the seamless observation of the complete factory environment to detect all dynamic changes. This global concept offers many advantages: the on-carrier sensors can be kept simple and cheap, driving paths can be early modified by the navigators and the actual detected carrier positions can be used for control purposes. The proposed system has a multi layer architecture consisting of the sensor layer, the low-level and high-level control layer. The sensor layer is responsible for the acquisition of environment data and low-level data analysis. The task of the low-level control layer is the improvement of the analysis data, using as much as possible a priori information. The high-level control layer executes the coordination of different observation tasks, using control rules which are activated by the results of the low-level control layer.

Library sharing system

Abstract: PURPOSE:To share a library as much as possible and to reduce cost for sharing and developing the library to plural operating systems (OS) by forming a standard interface at an upper layer part while dividing the library into upper and lower layer parts concerning the library sharing system to share the library. CONSTITUTION:A library 2 is divided into an upper layer part 3 and a lower layer part 4, the upper layer part 3 called from an application program 1 executes a prescribed processing and calls the lower layer part 4 as a standard interface 7, and the called lower layer part 4 calls an OS 5 as an OS interface 8 corresponding to the OS 5.

Neural network and its learning method

Abstract: PURPOSE:To provide the neural network which can learn even a linearly unseparable pattern and the learning method for that by conquering the problem that a perceptron can not linearly be separated and the problem that the local minimum point of back propagation is reached CONSTITUTION:In the neural network which has an input layer 1, an intermediate layer 2, and an output layer 3, a neural network model of low order consists of neurons of the intermediate layer 2 and input layer 1 which learn plural linearly separable patterns and a neural network model of high order consists of neurons of the intermediate layer 2 and output layer 3 which provide linearly unseparable patterns by integrating those linear separable patterns

An approach toward mechanization of acquiring reusable software components

Abstract: This paper gives a method for acquiring reusable software components automatically. Two categories of resources are distinguished, i.e., reusable patterns and reusable building blocks. It employs explanation--based learning method to analyse the concrete software design and generate its structural description, which, combined with semantic constraints, constitutes the reusable patterns. In order to obtain reusable building blocks, it applies both deduction and induction to construct the functional description of code fragment hierarchically, which , associated with code , is truly reusable in some sense. The ability of the system is enhanced with the acquisition of these reusable components and the scope of solvable problems is enlarged by employing them.

The Dynamically Reconfigurable Assembly System: Implementation Issues

Abstract: Conventional Automated Manufacturing Systems Development efforts generally result in a system tailored to the process and production volume requirements of a specific product family. Product-, process-, and device-based requirements are inextricably woven into one monolithic system which is very vulnerable to changes in either products, processes or production devices. In the new production environment, automated production systems must be able to rapidly adapt to new products and process, and to rapidly reconfigure and grow. The Automation & Robotics Research Institute is exploring new concepts in a system called the Dynamically Reconfigurable Assembly System. A key architectural principle used in the design and implementation of this system is the use of layered architectures. The physical system, the information and the control architectures are divided into three layers: Product, Process and Base System. Each layer in the hierarchy provides services to the layer above, while using the services of the layer below. This structure permits any layer in the system to be reconfigured without impacting the services provided to the adjacent layer. This paper will discuss some of the issues involved in the implementation of the physical and control architectures of this system.

Software structure for data processing system

Abstract: PURPOSE: To provide a software structure for the data processing system for real-time simulation which uses a large-scale software like, specially, a communication system. CONSTITUTION: The software structure is equipped with a service layer SE equipped with software function module of a system or function modules of application that do not include elements depending upon the hardware architecture of the system and a system layer SY equipped with the software function modules of the system. The system layer SY delivers messages between the software function modules of the application, manages the distribution and copy of the function modules of the application in the system, specifies and manages the reference to transaction processes between the function modules of the application, and allocates and protect the function module shooting of the application.