摘要:

AbstractSimple serial synchronized (SSS) multistage interconnection network (MIN) is a processor‐memory connection network that has a high performance/cost ratio, where the packet is inputted and switch synchronously in the MIN, which has a high pass‐through ratio and is composed of simple elements. This paper evaluates the effect of the hot spot contention and the effect of the synchronous bit‐serial (SBS) message combining in SSS‐MIN, by the theoretical analysis based on probability and simulation. In contrast to conventional MIN, there does not arise a complete tree saturation in SSS‐MIN, but an area, to which the access is difficult, is produced according to the relative position to the hot spot contention. From such a viewpoint, an analysis method for the pass‐through ratio is presented, which considers the position of the switching element to the hot spot. It is verified as a result of evaluation that the proposed method of analysis gives a result close to that of simulation, so long as the access to the hot spot and the connection network architecture stay within a practical range. It is also seen that the pass‐through ratio is deteriorated less in SSS‐MIN by the hot spot contention than in the conventional MIN, and the effect can be almost completely eliminated by the SBS message combining. When a multiprocessor system is actually constructed, performance deterioration due to hot spot contention is greater than in the case where only the pass‐through ratio is considered. This can also be eliminated almost completely by the SBS

ISSN:0882-1666    

DOI:10.1002/scj.4690260901

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1995

数据来源: WILEY

摘要:

AbstractThis paper proposes a superscalar architecture SARCH. An efficient mechanism for exploiting instruction‐level parallelism (ILP) is required to achieve high performance on nonnumerical applications. SARCH employs boosting for speculative execution to exploit large ILP, and a branch scheme to reduce branch penalty. Although these mechanisms are realized by simple hardware, code motion beyond basic blocks by compilers is necessary. A code scheduler is developed which performs global scheduling for the boosting and the delayed branch. The performance evaluation shows that the scheduled code achieves 1.52 time performance improvement over the original code, and 1.75x speed‐up over the scalar proces

ISSN:0882-1666    

DOI:10.1002/scj.4690260902

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1995

数据来源: WILEY

摘要:

AbstractThis paper proposes a method using the hypothetical reasoning for locating missing high cards in the game of bridge. Beginner knowledge for bidding and leading cards is represented by constraint logic programming language CHIP. Observation as to whether one leads with the king of hearts is given as a part of query. Then hypothesizing as to the possession of the ace of hearts is generated by abduction and integrity constraints are checked by consistency techniques in CHIP. A system based on beginners' knowledge is as powerful as an expert system used for the same purpose.

ISSN:0882-1666    

DOI:10.1002/scj.4690260903

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1995

数据来源: WILEY

摘要:

AbstractA new competitive learning approach for optimal vector quantizer is presented. First, it is shown that the original Competitive Learning (CL) algorithm can be derived from the two necessary conditions forming the basis of the traditional nonconnectionist vector quantizer (VQ) design algorithm called the LBG algorithm. It is then shown that the conventional conscience principle or equiprobable principle is not optimal from the standpoint of the minimization of the expected dis‐tortion. Next, a basic principle is introduced. Called the equidistortion principle (a necessary condition for optimal vector quantizers) it is derived by using Gersho's asymptotic theory. It is shown to be applicable to distributions consisting of disjoint clusters. Then a new competitive learning algorithm with a selection mechanism, called the competitive and selective learning (CSL) algorithm and based on the equidistortion principle, is proposed here. Since the selection mechanism enables the system to escape from local minima, the proposed algorithm can obtain better performance without a particular initialization procedure even when the input data clusters in a number of regions in the input vector space. In a synthetic one‐dimensional quantizer problem in which the optimum quantizer can be numerically obtained, less than 2% of the minimum distortion has been found which had not been obtained previously by con‐ventional algor

ISSN:0882-1666    

DOI:10.1002/scj.4690260904

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1995

数据来源: WILEY

摘要:

AbstractA new learning method for simple recurrent networks (SRN) is proposed. The correspondence between SRNs and finite stage machines (FSM) is examined and then a learning method for FSMs is constructed based on the state‐minimization technique of FSMs. This method comprises three stages: (1) generating states according to input histories; (2) grouping states according to outputs and next states; and (3) combining states within the same groups and reconstructing state transitions. Then, three networks, each behaving in the same manner as each stage through learning, are considered. A unified model of these networks and its learning method are proposed. The effectiveness of the proposed model and the learning method are verified by experimental result

ISSN:0882-1666    

DOI:10.1002/scj.4690260905

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1995

数据来源: WILEY

摘要:

AbstractA new mapping network combining wavelets and neural networks is proposed. The proposed network has three major characteristics, namely, self‐construction of neural networks, partial retrieval of the approximate mapping, and efficient learning. The network learning algorithm can be divided roughly into the self‐construction process and the error minimization process. In the first process, hidden units are appended serially in such a way that the network structure sufficiently approximates the approximation target. Simultaneously, network parameters are updated by competitive learning. In the second process, network parameters are updated by a localized backpropagation algorithm until the desired approximation accuracy is attained. The effectiveness of the proposed network is demonstrated through computer simulati

ISSN:0882-1666    

DOI:10.1002/scj.4690260906

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1995

数据来源: WILEY

摘要:

AbstractA high‐speed recursive algorithm has been proposed for determining multiple thresholds for a gray level of an image based on edge and contour evaluation using a two‐dimensional histogram of the maximum and minimum values in a 2 × 2 window and providing that the edge map of the image is given [5]. Although it is possible in principle to determine both the edge and gray level thresholds simultaneously by expanding the previous method, this requires excessive computation (proportional to the cube of the gray levels).This paper proposes a method of simultaneously determining the thresholds for both the edge and gray level so that the coincidence between the edge and the contour is optimized. In this method, the edge intensity is defined as the difference between the maximum and minimum values in a window; conditions that each pixel in an image is a part of a contour and/or an edge are discussed; and a set of optimum thresholds is determined with a relatively small amount of computation (proportional to the square of the gray levels) by using only a 2‐D histogram of two gray levels of the images after the maximum and the minimum

ISSN:0882-1666    

DOI:10.1002/scj.4690260907

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1995

数据来源: WILEY

摘要:

AbstractThe authors have previously developed a computer code for obtaining the current distribution on the magnetic resonance imaging (MRI) antenna loaded by a male model with an arbitrary geometry and material properties by incorporating the impedance method into the method of Richmond. Using this method, RF magnetic field penetration in a biological body within an MRI antenna is discussed theoretically. It is found that the RF magnetic field produced by the current in the body does not cancel the incident RF magnetic field produced by the current on the MRI antenna, since the phase difference between these magnetic fields deviates from 180°

ISSN:0882-1666    

DOI:10.1002/scj.4690260908

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1995

数据来源: WILEY

摘要:

AbstractHuman‐motion recognition techniques so far present several problems, such as difficult knowledge representation and poor robustness. With the aim of achieving recognition of human motion in a robust way by using knowledge representation that is easy to be understood, this paper proposes a method utilizing a fuzzy associative memory system.In the proposed method, the fuzzy associative memory system embodies fuzzy inference rules which define motion patterns of characteristic states contained in the time‐series pattern representing the motion. In the fuzzy associative memory system, inference is carried out with state transition patterns as input, and motion recognition is performed.The inference in the fuzzy associative memory system has the following two characteristics: (1) since knowledge can be represented as inference rules, it is easy to be understood by humans; (2) even in ambiguous situations or when part of the input information is missing, the information can be restored, or the degree of ambiguity can be controlled at inference time, etc., by the bidirectional processing of the conditional and action parts of the rules, so that appropriate inference results are obtained.In the proposed method, the time‐series motion patterns are converted into state transition patterns, and by using an associative inference system with the aforementionedcharacteristics, it is possible to perform robust recognition. The effectiveness of the proposed method is demonstrated through on‐line experimental results for the motion of nonspecific indi

ISSN:0882-1666    

DOI:10.1002/scj.4690260909

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1995

数据来源: WILEY