PRAND: GPU accelerated parallel random number generation library: Using most reliable algorithms and applying parallelism of modern GPUs and CPUs

The library PRAND for pseudorandom number generation for modern CPUs and GPUs is presented. It contains both single-threaded and multi-threaded realizations of a number of modern and most reliable generators recently proposed and studied in Barash (2011), Matsumoto and Tishimura (1998), L’Ecuyer (1999,1999), Barash and Shchur (2006) and the efficient SIMD realizations proposed in Barash and Shchur (2011). One of the useful features for using PRAND in parallel simulations is the ability to initialize up to 10¹⁹${10}^{19}$ independent streams. Using massive parallelism of modern GPUs and SIMD parallelism of modern CPUs substantially improves performance of the generators.     

Application of an improved control-variate scheme to local neoclassical transport simulations

A new radially local neoclassical transport code is developed based on the radially global two-weight δf$\delta f$ Monte Carlo code, FORTEC-3D (Satake et al., 2008). In the collisional two-weight δf$\delta f$ method, the variance of weight increases in time due to the so-called weight spreading, which leads to an increasing numerical noise in long time simulations. A new improved control-variate scheme was proposed by Kleiber et al. (2011) to reduce the variance. We investigate the effectiveness and validity of the new control-variate scheme for a practical collisional transport problem in a plasma described by the drift kinetic equation. It is demonstrated that the new scheme reduces the variance of weight by approximately 75% in an axisymmetric magnetic field configuration. The burst-like behavior of a transport observable caused by a large numerical noise can be avoided. As a result, the time evolution of a transport observable can be successfully smoothed.     

A heuristic biomarker selection approach based on professional tennis player ranking strategy

Extracting significant features from high-dimension and small sample size biological data is a challenging problem. Recently, Micha? Draminski proposed the Monte Carlo feature selection (MC) algorithm, which was able to search over large feature spaces and achieved better classification accuracies. However in MC the information of feature rank variations is not utilized and the ranks of features are not dynamically updated. Here, we propose a novel feature selection algorithm which integrates the ideas of the professional tennis players ranking, such as seed players and dynamic ranking, into Monte Carlo simulation. Seed players make the feature selection game more competitive and selective. The strategy of dynamic ranking ensures that it is always the current best players to take part in each competition. The proposed algorithm is tested on 8 biological datasets. Results demonstrate that the proposed method is computationally efficient, stable and has favorable performance in classification.     

Nonlinear time series forecasting with Bayesian neural networks

The Bayesian learning provides a natural way to model the nonlinear structure as the artificial neural networks due to their capability to cope with the model complexity. In this paper, an evolutionary Monte Carlo (MC) algorithm is proposed to train the Bayesian neural networks (BNNs) for the time series forecasting. This approach called as Genetic MC is based on Gaussian approximation with recursive hyperparameter. Genetic MC integrates MC simulations with the genetic algorithms and the fuzzy membership functions. In the implementations, Genetic MC is compared with the traditional neural networks and time series techniques in terms of their forecasting performances over the weekly sales of a Finance Magazine.     

基于Monte Carlo随机模拟的综合交通网络布局评价与决策方法

针对综合交通网络评价指标权重及属性值具有主观性和随机性的特点,提出了基于模拟运算的布局规划方案排序选优的群体随机决策方法.仿真生成满足集结的多个专家对指标重要性偏好排序统计分布的权重,同时考虑交通需求的不确定性对指标值的影响,结合客观熵权计算方案的综合评价值,由多次模拟得到的排序优势度确定方案的优劣差异.根据设计的仿真流程通过算例说明了方法应用的有效性,评价中考虑了主客观因素及随机性特征,可以为网络布局提供科学决策依据.     

SIMULATION AND ESTIMATION OF LONG MEMORY CONTINUOUS TIME MODELS

Abstract. Some general properties of long memory continuous time processes are recalled or proved. Methods of simulation are studied. A comparison with the usual discrete time autoregressive fractionally integrated moving-average filter is made and illustrations are provided. Then, two methods of estimation of the parameters of such a model from a discrete sample are studied, both theoretically and empirically, with Monte Carlo experiments.     

绝氧条件下β-Bi2Mo2O9晶相上丙烯氨氧化反应的Monte Carlo模拟研究

A random walk Monte Carlo （RWMC） simulation model of catalytic particle was established on the basis of the structures of bismuth molybdate catalysts and mechanisms of catalytic reactions with propylene selective oxidation and ammoxidation. The simulation results show that rationality of the RWMC model is proved on the basis of pulse experimental data. One of the most remarkable factors affecting catalytic behavior is the transfer of bulk lattice oxygen, which decides the rate of ammonia-consuming and propylene-consuming. The selectivity of main products reaches the maximum after the reduction of catalysts to a certain degree. It is inferred that catalytic performance improves greatly if the ratio of capacity for dehydrogenation from adsorbed propylene molecule on catalytically active site of molybdenum metal-imido group （Mo=NH） to that on catalytically active site of molybdenum metal-oxo group （Mo=O） becomes much higher.     

Bayesian Model Uncertainty In Smooth Transition Autoregressions

Abstract. In this paper, we propose a fully Bayesian approach to the special class of nonlinear time‐series models called the logistic smooth transition autoregressive (LSTAR) model. Initially, a Gibbs sampler is proposed for the LSTAR where the lag length, k, is kept fixed. Then, uncertainty about k is taken into account and a novel reversible jump Markov Chain Monte Carlo (RJMCMC) algorithm is proposed. We compared our RJMCMC algorithm with well‐known information criteria, such as the Akaikes? information criteria, the Bayesian information criteria (BIC) and the deviance information criteria. Our methodology is extensively studied against simulated and real‐time series.     

Joint Determination of the State Dimension and Autoregressive Order for Models with Markov Regime Switching

Abstract. This paper is concerned with the problem of joint determination of the state dimension and autoregressive order of models with Markov‐switching parameters. A model selection procedure is proposed which is based on optimization of complexity‐penalized likelihood criteria. The efficacy of the procedure is evaluated by means of Monte Carlo experiments.     

Unsteady radiative heat transfer within a suspension of ZnO particles undergoing thermal dissociation

An emitting, absorbing, and anisotropically scattering plain medium containing a suspension of ZnO particles is considered, in which the particles are directly exposed to high-flux irradiation and undergo shrinkage during their endothermic dissociation into Zn(g) and O₂ at above 2100 K. The unsteady energy equation that links the rate of radiative heat transfer to the rate of the chemical reaction is formulated and solved numerically by the finite volume technique and the explicit Euler time-integration scheme. The path-length Monte Carlo method is applied for modeling the radiative transfer within the suspension using the absorption/scattering coefficients and the scattering phase function obtained from the Mie theory. It is found that the particle suspension can be heated rapidly from its initial 300 K to over 1800 K in less than 0.1 s, resulting in a more uniform temperature profile as the reaction progresses, particles shrink, and the suspension becomes optically thinner. The chemical conversion increases with decreasing initial particle diameter and volume fraction due to the efficient radiative absorption.