Search

Simulations are more and more expensive in time computation and in number of models, so it is important to offer an effective solution to optimize distributed DEVS simulations in a pessimistic context. That is why, we propose to implement an optimization approach of distributed simulations by restructuration of the DEVS model hierarchy using graph partitioning tools. Indeed, the atomic models distribution can be seen like a partitioning graph problem, where the objective is to create a balanced model partition, with a minimal edge cut. However, the graph as designed does not reflect the dynamics of atomic models. The weighting of graph edges is based on a learning approach of each dynamic using hidden Markov models, with several adaptations based on the DEVS constraints. This article presents all of the process to obtaining an optimization of distributed simulation times. [ABSTRACT FROM AUTHOR]

Concerning the problem of large rotating machinery with non-stationary state like wind turbine, this research mainly makes an emphasis on the method of state deterioration recognition based on bi-spectrum entropy and HMM (Hidden Markov Model). Firstly, the true signal such as low-speed start vibration signals of rotor test rig in the normal state and a plurality of imbalance deterioration degrees are collected. Bi-spectrum is applied to obtain the fault feature from the vibration signals mixed with a complex background noise. On the basis of bi-spectrum analysis, a bi-spectrum entropy algorithm is derived under the condition of subspace distribution probability, and the HMM for the fault pattern recognition is established by using the bi-spectrum entropy feature as input. This method is verified by successfully recognizing four state deterioration degrees. Finally, the method is applied to recognize the imbalance deterioration degree of wind turbine with the type of SL1500/82 and equipment actual working condition verified the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Multiscale models are designed to handle problems with different length scales and time scales in a suitable and efficient manner. Such problems include inelastic deformation or failure of materials. In particular, hierarchical multiscale methods are computationally powerful as no direct coupling between the scales is given. This paper proposes a hierarchical two-scale setting appropriate for isothermal quasi-static problems: a macroscale treated by continuum mechanics and the finite element method and a microscale modelled by a canonical ensemble of statistical mechanics solved with molecular dynamics. This model will be implemented into the framework of the heterogeneous multiscale method. The focus is laid on an efficient coupling of the macro- and micro-solvers. An iterative solution algorithm presents the macroscopic solver, which invokes for each iteration an atomistic computation. As the microscopic computation is considered to be very time consuming, two optimisation strategies are proposed. Firstly, the macroscopic solver is chosen to reduce the number of required iterations to a minimum. Secondly, the number of time steps used for the time average on the microscale will be increased with each iteration. As a result, the molecular dynamics cell will be allowed to reach its state of thermodynamic equilibrium only in the last macroscopic iteration step. In the preceding iteration steps, the molecular dynamics cell will reach a state close to equilibrium by using considerably fewer microscopic time steps. This adapted number of microsteps will result in an accelerated algorithm (aFE-MD-HMM) obtaining the same accuracy of results at significantly reduced computational cost. Numerical examples demonstrate the performance of the proposed scheme. [ABSTRACT FROM AUTHOR]

Numerical geodynamo simulations with parameters close to an Earth-like regime would be of great interest for understanding the dynamics of the Earth's liquid outer core and the associated geomagnetic field. Such simulations are far too computationally demanding owing to the large range in spatiotemporal scales. This paper explores the application of a multi-scale timestepping method to an asymptotic model for the generation of magnetic field in the fluid outer core of the Earth. The method is based on the heterogeneous multiscale modelling (HMM) strategy, which incorporates scale separation and utilises several integrating models for the fast and slow fields. Quantitative comparisons between the multi-scale simulations and direct solution of the asymptotic model in the limit of rapid rotation and low Ekman number are performed. The multi-scale method accurately captures the varying temporal and spatial dynamics of the mean magnetic field at lower computational costs compared to the direct solution of the asymptotic model. [ABSTRACT FROM AUTHOR]

Abstract: The contourlet transform is an emerging multiscale multidirection image processing technique. It effectively represents smooth curvature details typical of natural images, overcoming a major drawback of the 2-D wavelet transform. Previously, we developed a contourlet image model, that is, the contourlet contextual hidden Markov model (C-CHMM). In this paper, we further develop a multiscale texture segmentation technique based on the C-CHMM. The segmentation method combines a model comparison approach with a multiscale fusion and a neighbor combination process. It also features a neighborhood selection scheme based on smoothed context maps, for both model estimation and neighbor combination. Through a series of segmentation experiments, we examine the effectiveness of the C-CHMM in comparison with closely related models. We also investigate how different context designs affect the segmentation performance. Moreover, we show that the C-CHMM based technique provides improved accuracy in segmenting texture patterns of diversified nature, as compared with popular methods such as the HMTseg and the JMCMS. All these simulation experiments demonstrate the great potential of the C-CHMM for image analysis applications. [Copyright &y& Elsevier]

Abstract: Hidden Markov Models (HMMs) are practical tools which provide probabilistic base for protein secondary structure prediction. In these models, usually, only the information of the left hand side of an amino acid is considered. Accordingly, these models seem to be inefficient with respect to long range correlations. In this work we discuss a Segmental Semi Markov Model (SSMM) in which the information of both sides of amino acids are considered. It is assumed and seemed reasonable that the information on both sides of an amino acid can provide a suitable tool for measuring dependencies. We consider these dependencies by dividing them into shorter dependencies. Each of these dependency models can be applied for estimating the probability of segments in structural classes. Several conditional probabilities concerning dependency of an amino acid to the residues appeared on its both sides are considered. Based on these conditional probabilities a weighted model is obtained to calculate the probability of each segment in a structure. This results in 2.27% increase in prediction accuracy in comparison with the ordinary Segmental Semi Markov Models, SSMMs. We also compare the performance of our model with that of the Segmental Semi Markov Model introduced by Schmidler et al. [C.S. Schmidler, J.S. Liu, D.L. Brutlag, Bayesian segmentation of protein secondary structure, J. Comp. Biol. 7(1/2) (2000) 233–248]. The calculations show that the overall prediction accuracy of our model is higher than the SSMM introduced by Schmidler. [Copyright &y& Elsevier]

Abstract: A procedure is proposed for computing the autocovariances and the ARMA representations of the squares, and higher-order powers, of Markov-switching GARCH models. It is shown that many interesting subclasses of the general model can be discriminated in view of their autocovariance structures. Explicit derivation of the autocovariances allows for parameter estimation in the general model, via a GMM procedure. It can also be used to determine how many ARMA representations are needed to identify the Markov-switching GARCH parameters. A Monte Carlo study and an application to the Standard & Poor index are presented. [Copyright &y& Elsevier]