摘要:

AbstractAbstract This paper presents a Petri net model, called the interpreted coloured Petri net, which allows the verification of parallel systems specified with the ISO language Estelle. The main contribution of this model is to take into account parallel behaviours with strong synchronism constraints. Such behaviours are often encountered in real-time systems in which parallel execution is controlled by a clock tip. Our model allows the expression of simultaneous executions as it can be specified in Estelle. Moreover, this model is the basis of a protocol verification tool called Petristelle.

ISSN:0228-6203    

DOI:10.1080/02286203.1996.11760279

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

AbstractSelf-affine and piecewise self-affine IFS fractal models are used in this paper to model several different types of discrete sequences. The parameters of such models are determined according to an optimizar tion criterion. However, the general optimization problem is quite complex, and therefore some constraints are introduced. The best tradeoff between overall performance and computational complexity is found. The optimal estimation of the fractal models parameters is obtained by means of genetic algorithms, and a very good convergence to the global minimum is obtained with a proper tuning of the algorithm. A comparison with suboptimal algorithms is reported. Several types of discrete sequences are modelled and the performance results are described. The genetic optimization algorithm behaves quite well in comparison to suboptimal approaches in terms of both performance and computational complexity.

ISSN:0228-6203    

DOI:10.1080/02286203.1996.11760280

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

AbstractHuffman coding is known for its ability to generate minimumredundancy codes given a set of nonnegative weights. Associative memory is one of the most widely used tools in parallel processing because of its capability of parallel search and comparisons. In this paper we develop a pair of implementations for static and dynamic Huffman coding based on a sequential representation using associative memory. The efficiency of the proposed implementations is evaluated and the simulation results are presented.

ISSN:0228-6203    

DOI:10.1080/02286203.1996.11760281

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

AbstractNeural networks are powerful computational tools and have been applied in various ways. In this work a neural network is used to solve a pattern classification problem encountered in biochemistry. One of the major topics of research in molecular biology is the prediction of functional properties of biomacroinoleculcs from their sequence data. A radial basis function (RBF) network is presented for predicting the trausmembrane regions of membrane proteins. A comparison between the results of prediction by the RBF network and a popular conventional method along with real data show the superiority of this approach to prediction of transmembrane regions.

ISSN:0228-6203    

DOI:10.1080/02286203.1996.11760282

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

AbstractThe critical load of a frame is affected to a large extent by variations in the lengths and cross-sections of beams and columns and the strength of steel. Consequently, the deterministic critical load does not represent the true behaviour of the frame. In this paper, based on a rigorous probabilistic analysis, closed-form equations are derived between the coefficient of variation of the critical load of the frame and the corresponding dependent and independent variables. These are then checked with the results obtained using Monte Carlo simulation. The practical significance of these results vis-à-vis deterministic analysis is discussed.

ISSN:0228-6203    

DOI:10.1080/02286203.1996.11760283

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

AbstractIn most manufacturing environments, data representing an object are scattered over multiple databases and can be updated by multiple sources. The author refers to such an environment as a multidatabase multisource (MDMS) data environment. One of the most important issues in an MDMS data environment is how to choose a value for an object when there exist multiple records of values for it. We present a model, called tuneup-schema multidatabase, by which the most beneficial set of values for a manufacturing object is determined, and the manufacturing process is optimized.

ISSN:0228-6203    

DOI:10.1080/02286203.1996.11760284

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

AbstractA technique used for three-dimensional reconstruction of microscopic objects is based on linear processing of a collection of images taken at different focus positions, that is, acquired step by step while the objective moves along the microscope optical axis. As the relationship between the“object”I(x,y,z) and the collection of images (if these latter are enough) is known to be a convolution via the microscope point spread function, a deconvolution procedure yields I(x,y,z). Here I(x,y,z) is the local absoiption of the object. Based on the fact that reasonable approximations for the point spread function are available in the literature, the above technique seems to work satisfactorily, and has been applied by various researchers. In the present work, the method is analyzed from a strictly mathematical point of view. It Is shown that being the optical transfer function (OTF), that is, the Fourier transform of the point spread function, identically vanishing in an infinite region of the frequency domain, a deconvolution procedure is not feasible. If the z-coordinate refers to the optical axis, the OTF is actually zero outside a conical region in the frequency domain. It is also shown that what can be actually recovered are projections of the given object. To this end, the so-called slice theorem, or projection theorem, is used. More exactly, we prove that one can obtain all projections of the object within an angle of projection-with respect to the optical axis-not greater than the aperture angle of the image formation system, that is, the microscope. Examples of simulations and of processing of experimental images are reported.

ISSN:0228-6203    

DOI:10.1080/02286203.1996.11760285

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

AbstractThe authors propose a new algorithm to obtain and validate the standard singularly perturbed form for linear discrete-time systems. The technique used is based on Gerschgorin’s circle method and gives a good selection of the slow and fast subsystems for a system. The algorithm uses the iterative computation of block-diagonalization matrices that are available for continuous-time singularly perturbed systems.

ISSN:0228-6203    

DOI:10.1080/02286203.1996.11760286

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor