Bayesian Learning of Noisy Markov Decision Processes

We consider the inverse reinforcement learning problem, that is, the problem of learning from, and then predicting or mimicking a controller based on state/action data. We propose a statistical model for such data, derived from the structure of a Markov decision process. Adopting a Bayesian approach to inference, we show how latent variables of the model can be estimated, and how predictions about actions can be made, in a unified framework. A new Markov chain Monte Carlo (MCMC) sampler is devised for simulation from the posterior distribution. This step includes a parameter expansion step, which is shown to be essential for good convergence properties of the MCMC sampler. As an illustration, the method is applied to learning a human controller.

Modeling mode i fracture of bitumen films

The fracture behavior of thin films of bitumen in double cantilever beam (DCB) specimens was investigated over a wide range of temperature and loading rate conditions using finite-element analysis. The model includes a phenomenological model for the mechanical behavior of bitumen, implemented into a special-purpose finite-element user material subroutine, combined with a cohesive zone model (CZM) for simulating the fracture process. The finite-element model is validated against experimental results from laboratory tests of DCB specimens by comparing measured and predicted load-line deflection histories and fracture energy release rates. Computer simulation results agreed well with experimental data of DCB joints containing bitumen films in terms of peak stress, fracture toughness, and stress-strain history response. The predicted "normalized toughness," G=2h, was found to increase in a power-law manner with effective temperaturecompensated strain rate in the ductile region as previously observed experimentally. In the brittle regime, G=2h is virtually constant. The model successfully captured the ductile and brittle failure behavior of bitumen films in opening mode (tension) for stable crack growth conditions. © 2013 American Society of Civil Engineers.

The influence of internal electric fields on the transition energy of InGaN/gaN quantum well

We present a new way to meet the amount of strain relaxation in an InGaN quantum well layer grown on relaxed GaN by calculating and measuring its internal field. With perturbation theory, we also calculate the transition energy of InGaN/GaN SQWs as affected by internal fields. The newly reported experimental data by Graham et al. fit our calculations well on the assumption that the InGaN well layer suffered a 20% strain relaxation, we discuss the differences between our calculated results and the experimental data. Our calculation suggests that with the increase of indium mole fraction in the InGaN/GaN quantum well, the effect of polarization fields on the luminescence of the quantum well will increase. Moreover, our calculation also suggests that an increase in the quantum well width by only one monolayer can result in a large reduction in the transition energy. (c) 2006 Elsevier B.V. All rights reserved.

Electronic structures of wurtzite ZnO and ZnO/MgZnO quantum well

The band structures of wurtzite ZnO are calculated using the empirical pseudopotential method (EPM). The 8 parameters of the Zn and O atom pesudopotential form factors with Schluter's formula are obtained. The effective mass parameters are extracted by using k.p Hamiltonian to fit the EPM results. The calculated band edge energies (E-g, E-A, E-B, and E-C) at Gamma point are in good agreement with experimental results. The ordering of ZnO at the top of valence band is found to be A(Gamma(7))-B(Gamma(9))-C(Gamma(7)) due to a negative spin-orbit (SO) splitting. Based on the band parameters obtained, the valence hole subbands of wurzite ZnO/MgxZn1-xO tensile-strained quantum wells (QWs) with different well widths and Mg compositions are calculated using 6-band k.p method. (c) 2005 Elsevier B.V. All rights reserved.

Effects and numerical analysis of argon gas flow on the oxygen concentration in Czochralski silicon single crystal growth

Argon gas, as a protective environment and carrier of latent heat, has an important effect on the temperature distribution in crystals and melts. Numeric simulation is a potent tool for solving engineering problems. In this paper, the relationship between argon gas flow and oxygen concentration in silicon crystals was studied systematically. A flowing stream of argon gas is described by numeric simulation for the first time. Therefore, the results of experiments can be explained, and the optimum argon flow with the lowest oxygen concentration can be achieved. (C) 2002 Elsevier Science B.V. All rights reserved.

A geometrical model of GaN morphology in initial growth stage

Based on morphology observed by atomic force microscopy, a geometrical model was proposed in order to explain the statistical results obtained from morphology observation on GaN in initial growth stage. Four parameters were introduced to describe the morphology characteristics in this model. Least-square fitting of height distribution was performed. The height distribution derived from the model agreed well with that obtained from experimental records. It was also found that the model should be further advanced to understand the growth of GaN in initial growth stage. (C) 2002 Elsevier Science BY. All rights reserved.

Quasi-thermo dynamic analysis of MOVPE growth of GaxAlyIn1-x-yN

A quasi-thermodynamic model of metalorganic vapor phase epitaxy (MOVPE) growth of GaxAlyIn1-x-yN alloys has been proposed. In view of the complex growth behavior of GaxAlyIn1-x-yN, we focus our attention on the galliumrich quaternary alloys that are lattice matched to GaN, In0.15Ga0.85N or Al0.15Ga0.85N, which are widely used in the GaN-based optoelectronic devices. The relationship between GaAlInN alloy composition and input molar ratio of group III metalorganic compounds at various growth conditions has been calculated. The influence of growth temperature, nitrogen fraction in the carrier gas, input partial pressure of group III metalorganics, reactor pressure, V/III ratio and the decomposition rate of ammonia on the composition of deposited alloys are studied systematically. Based on these calculated results, we can find out the appropriate growth conditions for the MOVPE growth of GaxAlyIn1-x-yN alloy lattice matched to GaN, In0.15Ga0.85N or Al0.15Ga0.85N. (C) 2002 Elsevier Science B.V. All rights reserved.

Electronic structures of wurtzite ZnO and ZnO/MgZnO quantum well

The band structures of wurtzite ZnO are calculated using the empirical pseudopotential method (EPM). The 8 parameters of the Zn and O atom pesudopotential form factors with Schluter's formula are obtained. The effective mass parameters are extracted by using k.p Hamiltonian to fit the EPM results. The calculated band edge energies (E-g, E-A, E-B, and E-C) at Gamma point are in good agreement with experimental results. The ordering of ZnO at the top of valence band is found to be A(Gamma(7))-B(Gamma(9))-C(Gamma(7)) due to a negative spin-orbit (SO) splitting. Based on the band parameters obtained, the valence hole subbands of wurzite ZnO/MgxZn1-xO tensile-strained quantum wells (QWs) with different well widths and Mg compositions are calculated using 6-band k.p method. (c) 2005 Elsevier B.V. All rights reserved.

Phase reconstruction from three interferograms based on integral of phase gradient

A novel algorithm of phase reconstruction based on the integral of phase gradient is presented. The algorithm directly derives two real-valued partial derivatives from three phase-shifted interferograms. Through integrating the phase derivatives, the desired phase is reconstructed. During the phase reconstruction process, there is no need for an extra rewrapping manipulation to ensure values of the phase derivatives lie in the interval [-pi, pi] as before, thus this algorithm can prevent error or distortion brought about by the phase unwrapping operation. Additionally, this algorithm is fast and easy to implement, and insensitive to the nonuniformity of the intensity distribution of the interferogram. The feasibility of the algorithm is demonstrated by both computer simulation and experiment.

数控光学成形的计算机模拟及工艺实验

本文详细地介绍了以控制磨头驻留时间为基础的CCOS的基本原理、实现方法以及计算机模拟软件和工艺实验。在基本原理的基础上，着重研究解决了CCOS计算机模拟当中的两个关键问题“边缘效应”和计算效率的提高。首先，在分析计算磨头工作函数在边缘区域的变化以及边缘区域材料去除量变化规律的基础上，得到磨头在边缘区域驻留时间函数的修正函数。其次，对原有的模拟计算算法进行了改进，改进后算法的计算速度比原来快了十几倍从而大大提高了加工及检测的效率。另外还进行了CCOS的工艺实验研究，确定了CCOS抛光阶段的加工效率约为6×10~(-4)mm~3/sec为预计绝对加工时间提供了依据。

ESTIMATION OF THE STRAIN STATE OF GEXSI1-X/SI STRAINED-LAYER SUPERLATTICES BY DOUBLE-CRYSTAL X-RAY-DIFFRACTION

20-period strained-layer superlattices of nominal composition and width Ge0.2Si0.8 (5 nm)/Si(25 nm) and Ge0.5Si0.5 (5 nm)/Si(25 nm) were studied by double-crystal X-ray diffraction. The Ge content x was determined by computer simulation of the diffraction features from the superlattice. This method is shown to be independent of the relaxation of the superlattice. Alternatively, x can be obtained from the measured difference DELTAa/a in lattice spacing perpendicular to the growth plane. It is sensitive to the relaxation. Comparing the results obtained in these two different ways, information about the relaxation of the superlattices can be obtained.

THE DEPENDENCE OF THE INVERSION LAYER THICKNESS ON THE FILM THICKNESS IN THIN-FILM SOI STRUCTURES

The dependence of the inversion-layer thickness on the film thickness in thin-film SOI structure is analyzed theoretically by using computer simulation. A new concept and parameter, the critical thickness of thin film all-bulk inversion, is introduced for the design of thin-film MOS/SOI devices. It is necessary to select the film thickness T(s1) close to the all-bulk strong inversion critical thickness in order to get high-speed and high-power operation of ultra-thin film MOS/SOI devices.

A COMMON-CAVITY 2-SECTION INGAASP/INP BISTABLE LASER WITH A LOW OPTICAL SWITCHING POWER

The characteristics of the steady-state and the transient response to external light excitation of a common-cavity two-section (CCTS) bistable semiconductor laser is investigated. The results on the relation of light output versus light input, the wavelength match, optical amplification and optical switching are presented. Experimental results are compared to the results of a computer simulation.

Study of GSMBE growth and characteristics of high-quality strained In0.63Ga0.37As/InP quantum wells

High-quality compressively strained In0.63Ga0.37As/InP quantum wells with different well widths (1-11 nm) have been grown coherently on InP substrates using a home-made gas source molecular beam epitaxy (GSMBE) system. The indium composition in the wells of the sample was determined by means of high-resolution X-ray diffraction and its computer simulation. it is found that the exciton transition energies determined by photoluminescence (PL) at 10 K are in good agreement with those calculated using a deformation potential model. Sharp and intense peaks for each well can be well resolved in the 10 K PL spectra. For wells narrower than 4 nm, the line width of the PL peaks are smaller than the theoretical values of the line-width broadening due to 1 hit interface fluctuation, showing that the interface fluctuation of our sample is within 1 ML. For wells of 7 and 9 nm, the PL peak widths are as low as 4.5 meV.

Analysis of orthogonal memory patterns

It is well known that the storage capacity may be large if all memory patterns are orthogonal to each other. In this paper, a clear description is given about the relation between the dimension N and the maximal number of orthogonal vectors with components +/-1, and also the conception of attractive index is proposed to estimate the basin of attraction. Theoretic analysis and computer simulation show that each memory pattern's basin of attraction contains at least one Hamming ball when the storage capacity is less than 0.33N which is better than usual 0.15 N.