A new kernel-projective statistical estimator in the Monte Carlo method

2020 ◽  
Vol 35 (6) ◽  
pp. 341-353
Author(s):  
Gennady A. Mikhailov ◽  
Natalya V. Tracheva ◽  
Sergey A. Ukhinov

AbstractIn the present paper, we propose a new combined kernel-projective statistical estimator of the two-dimensional distribution density, where the first ‘main’ variable is processed with the kernel estimator, and the second one is processed with the projective estimator for the conditional distribution density. In this case, statistically estimated coefficients of some orthogonal expansion of the conditional distribution density are used for each ‘kernel’ interval defined by a micro-sample. The root-mean-square optimization of such an estimator is performed under the assumptions concerning the convergence rate of the used orthogonal expansion. The numerical study of the constructed estimator is implemented for angular distributions of the radiation flux forward-scattered and backscattered by a layer of matter. A comparative analysis of the results is performed for molecular and aerosol scattering.

2015 ◽  
Vol 2015 ◽  
pp. 1-8 ◽  
Author(s):  
Irina Stanciu

The geometric layout is the key factor for enhancing the efficiency of the fluid mixing in passive micromixers. Therefore, by adjusting the geometric design and by controlling the geometric parameters, one can enhance the mixing process. However, through any fabrication process, the geometric parameters present slight, inherent variation from the designed values than might affect the performance of the micromixer. This paper proposes a numerical study on the influence of the unavoidable geometric tolerances on the mixing efficiency in passive micromixers. A probabilistic simulation model, based on the Monte Carlo method, is developed and implemented for this purpose. An uncertainty simulation model shows that significant deviations from the deterministic design can appear due to small variations in the geometric parameters values and demonstrates how a more realistic mixing performance can be estimated.


2005 ◽  
Vol 295-296 ◽  
pp. 3-8 ◽  
Author(s):  
Taeho Ha ◽  
Takashi Miyoshi ◽  
Yasuhiro Takaya ◽  
S. Takahashi

Laser scattering characteristics of typical CMP-induced defects such as particles and microscratches with the size of sub-micron order are investigated using a developed automated scattered light measurement system. The measurement system has an ability to detect three-dimensional distribution of scattered light from the defects with high sensitivity. The angular distributions of scattered light from the standard PSL (Polystyrene latex) spheres and microscratches reveal that scattering characteristics of microscratches are quite different from those of Particles. The scattered light from the PSL sphere is detected mainly ahead. In contrast, the scattered light from the microscratches only exists in the direction perpendicular to its length dimension at an oblique incidence maintaining the sheet-shaped pattern, even if its orientation relative to the incident direction is changed. Optical arrangement for defect detection and classification is suggested based on the experimental results.


Author(s):  
С.А. Желтов ◽  
Н.Н. Чупятов

Процесс нанесения покрытий на поверхности является одним из важнейших при нанесении защитных слоев. Особенно актуальным является метод молекулярно-лучевой эпитаксии, который применяют, например, для нанесения покрытий металлов и полимеров на металлические поверхности. Однако при производстве установок необходимо провести много предварительных экспериментов для получения максимально однородных напыленных поверхностей: размеры кювет, уровень расплава в кювете, расстояния до напыляемой поверхности и расположения системы кювет. Упрощением проведения этих реальных экспериментов является метод Монте-Карло. Моделируя компьютерные эксперименты с предполагаемыми параметрами установки, можно быстро получить необходимый результат с высокой точностью. В каждом компьютерном эксперименте разыгрывался вылет N = 108 частиц (атомов, молекул), что обеспечивало получение всех результатов с высокой точностью. Было установлено, что результаты расчетов не зависят от масс частиц. Анализ полученных данных позволил установить, что частицы, испарявшиеся с поверхности расплава по равновероятному закону и по закону косинуса, имеют близкие распределения, отличающиеся не более чем на 10%. Увеличение расстояния от верхнего края пластины до напыляемой плоскости с расстояния G = 0,01 м до расстояния G = 0,05 м приводит к тому, что плотность распределения для первого случая, имеющая значительный максимум около середины кюветы, становится практически равномерным распределением во втором случае. Применение метода Монте-Карло для процессов напыления в вакууме эффективно использовать не только для плоских поверхностей, но и на поверхности произвольных форм, например, цилиндрических. The coating process on surfaces is one of the most important in the application of protective layers. Particularly relevant is the method of molecular beam epitaxy, which is used, for example, for the deposition of coatings of metals and polymers on metal surfaces. However, in the manufacture of installations, it is necessary to carry out many preliminary experiments to obtain the most homogeneous sprayed surfaces: the dimensions of the cuvettes, the level of the melt in the cuvette, the distance to the sprayed surface and the location of the cuvette system. A simplification of these real-world experiments is the Monte Carlo method. By simulating computer experiments with the expected parameters of the installation, it is possible to quickly obtain the required result with high accuracy. In each computer experiment, the emission of N = 108 particles (atoms, molecules) was played, which ensured obtaining all results with high accuracy. It was found that the calculation results do not depend on the particle masses. An analysis of the data obtained made it possible to establish that the particles evaporated from the surface of the melt according to the equiprobable law and according to the cosine law have similar distributions differing by no more than 10%. An increase in the distance from the upper edge of the plate to the sprayed plane from a distance of G = 0.01 m to a distance of G = 0.05 m leads to the fact that the distribution density for the first case, which has a significant maximum near the middle of the cell, becomes an almost uniform distribution in the second case. ... The use of the Monte Carlo method for vacuum deposition processes can be effectively used not only for flat surfaces, but also on surfaces of arbitrary shapes, for example, cylindrical ones.


2011 ◽  
Vol 45 (23) ◽  
pp. 2465-2473 ◽  
Author(s):  
Qin Zhang ◽  
Zhihua Pi ◽  
Mingxiang Chen ◽  
Xiaobing Luo ◽  
Ling Xu ◽  
...  

The effective thermal conductivity of silicone/phosphor composites is studied experimentally and numerically. Thermal conductivity measurements are conducted from 30°C to 150°C for the composites with phosphor volume fraction up to 40%. In the numerical study, a finite element model with empirical particle size distribution and random particle position is constructed using a probability density function and the Monte Carlo method, and the interfacial thermal resistance layer between phases also introduced in the model. The results indicate that when phosphor concentration is below 25 vol.%, the conductivity of the composite increases slightly with either phosphor volume fraction or temperature, and the Kapitza radius of the composite is 0.8 µm. When phosphor concentration is above 25 vol.%, the increase of conductivity correlates positively with phosphor volume fraction significantly but negatively with the temperature, and the Kapitza radius is 0.032 µm.


2015 ◽  
Vol 24 (01) ◽  
pp. 1550003 ◽  
Author(s):  
A. H. Al-Ghamdi ◽  
Awad A. Ibraheem ◽  
M. El-Azab Farid

The alpha (α) elastic scattering from different targets potential over the energy range 10–240 MeV has been analyzed in the framework of the single-folding (SF) optical model. Four targets are considered, namely, 24 Mg , 28 Si , 32 S and 40 Ca . The SF calculations for the real central part of the nuclear optical potential are performed by folding an effective α–α interaction with the α-cluster distribution density in the target nucleus. The imaginary part of the optical potential is expressed in the phenomenological Woods–Saxon (WS) form. The calculated angular distributions of the elastic scattering differential cross-section using the derived semimicroscopic potentials successfully reproduce 36 sets of data all over the measured angular ranges. The obtained results confirm the validity of the α-cluster structure of the considered nuclei. For the sake of comparison, the same sets of data are reanalyzed using microscopic double-folded optical potentials based upon the density-dependent Jeukenne–Lejeune–Mahaux (JLM) effective nucleon–nucleon interaction.


Author(s):  
Andrey Yu. Ambos ◽  
Gennady A. Mikhailov

AbstractA methodology of efficient homogenization of a radiation model for a stochastic media is developed. In this connection, ‘realistic’ computational models of random fields are constructed, implementations of those fields are close to continuous ones, and the conditional one-dimensional distribution in the nonempty part of the medium is quite natural. It is shown that the corresponding homogenized probability of the passage of a quantum particle is actually determined by the correlation scale and the occupancy of the medium and also by the above conditional distribution.


Author(s):  
KIRTEE K. KAMALJA

Reliability importance of components of various systems is an important part in the reliability study. In this paper Birnbaum-reliability importance (B-importance) of components of three popular consecutive-type systems as consecutive-k-out-of-n: F-system, m-consecutive-k-out-of-n: F system and r-within-consecutive-k-out-of-n: F system is studied. The B-importance is studied when the survival probabilities of components of the system are Markov dependent. This study is based on conditional distribution of number of occurrences of failure runs of length k and number of occurrences of scanning windows of length k containing r failures in the sequence of Markov Bernoulli trials. Simplified formulae for calculation of B-importance of each component of the three consecutive-type systems are developed in terms of survival probabilities. To demonstrate the simplicity of the derived results, numerical study is also included.


Sign in / Sign up

Export Citation Format

Share Document