Regression Density Estimation With Variational Methods and Stochastic Approximation

Regression density estimation is the problem of flexibly estimating a response distribution as a function of covariates. An important approach to regression density estimation uses finite mixture models and our article considers flexible mixtures of heteroscedastic regression (MHR) models where the response distribution is a normal mixture, with the component means, variances, and mixture weights all varying as a function of covariates. Our article develops fast variational approximation (VA) methods for inference. Our motivation is that alternative computationally intensive Markov chain Monte Carlo (MCMC) methods for fitting mixture models are difficult to apply when it is desired to fit models repeatedly in exploratory analysis and model choice. Our article makes three contributions. First, a VA for MHR models is described where the variational lower bound is in closed form. Second, the basic approximation can be improved by using stochastic approximation (SA) methods to perturb the initial solution to attain higher accuracy. Third, the advantages of our approach for model choice and evaluation compared with MCMC-based approaches are illustrated. These advantages are particularly compelling for time series data where repeated refitting for one-step-ahead prediction in model choice and diagnostics and in rolling-window computations is very common. Supplementary materials for the article are available online.     

Speeding up MCMC by Delayed Acceptance and Data Subsampling

The complexity of the Metropolis–Hastings (MH) algorithm arises from the requirement of a likelihood evaluation for the full dataset in each iteration. One solution has been proposed to speed up the algorithm by a delayed acceptance approach where the acceptance decision proceeds in two stages. In the first stage, an estimate of the likelihood based on a random subsample determines if it is likely that the draw will be accepted and, if so, the second stage uses the full data likelihood to decide upon final acceptance. Evaluating the full data likelihood is thus avoided for draws that are unlikely to be accepted. We propose a more precise likelihood estimator that incorporates auxiliary information about the full data likelihood while only operating on a sparse set of the data. We prove that the resulting delayed acceptance MH is more efficient. The caveat of this approach is that the full dataset needs to be evaluated in the second stage. We therefore propose to substitute this evaluation by an estimate and construct a state-dependent approximation thereof to use in the first stage. This results in an algorithm that (i) can use a smaller subsample m by leveraging on recent advances in Pseudo-Marginal MH (PMMH) and (ii) is provably within O(m^{? 2}) of the true posterior.     

Combining column generation and constraint programming to solve the tail assignment problem

Within the area of short term airline operational planning, Tail Assignment is the problem of assigning flight legs to individual identified aircraft while satisfying all operational constraints, and optimizing some objective function. In this article, we propose that Tail Assignment should be solved as part of both the short and the long term airline planning. We further present a hybrid column generation and constraint programming solution approach. This approach can be used to quickly produce solutions for operations management, and also to produce close-to-optimal solutions for long and mid term planning scenarios. We present computational results which illustrate the practical usefulness of the approach.     

Strong and coalition-proof political equilibria under plurality and runoff rule

We analyze the effects of voter coordination on the equilibrium of voting games. Specifically, we analyze the concepts of strong equilibrium and coalition-proof equilibrium for plurality rule and runoff rule elections. We characterize these equilibria in the three candidate case, and provide results for important special cases in the general multicandidate case.      

Nonlinear Isolator Dynamics at Finite Deformations: An Effective Hyperelastic,Fractional Derivative,Generalized Friction Model

In presenting a nonlinear dynamic model of a rubber vibrationisolator, the quasistatic and dynamic motion influences on theforce response are investigated within the time and frequencydomain. It is found that the dynamic stiffness at the frequency ofa harmonic displacement excitation, superimposed upon the longterm isolator response, is strongly dependent on staticprecompression, dynamic amplitude and frequency. The problems ofsimultaneously modelling the elastic, viscoelastic and frictionforces are removed by additively splitting them, modelling theelastic force response by a nonlinear, shape factor basedapproach, displaying results that agree with those of aneo-Hookean hyperelastic isolator at a long term precompression.The viscoelastic force is modeled by a fractional derivativeelement, while the friction force governs from a generalizedfriction element displaying a smoothed Coulomb force. A harmonicdisplacement excitation is shown to result in a force responsecontaining the excitation frequency and its every otherhigher-order harmonic, while using a linearized elastic forceresponse model, whereas all higher-order harmonics are present forthe fully nonlinear case. It is furthermore found that the dynamicstiffness magnitude increases with static precompression andfrequency, while decreasing with dynamic excitationamplitude – eventually increasing at the highest amplitudes due tononlinear elastic effects – with its loss angle displaying amaximum at an intermediate amplitude. Finally, the dynamicstiffness at a static precompression, using a linearized elasticforce response model, is shown to agree with the fully nonlinearmodel except at the highest dynamic amplitudes.     

Combination of schlieren and pulsed TV holography in the study of a high-speed flame jet

This paper presents studies of the propagation of a high-speed turbulent flame jet of an air/hydrogen gas mixture. The experimental results are recorded with the schlieren and the pulsed TV holography method. These methods are compared and combined to benefit from the advantages of each of them. Abel inversion has been used to achieve three-dimensional information i.e. refractive index distributions. Evaluation, calibration and Abel inversion algorithms are described together with experimental results. The results obtained from the different techniques show remarkable similarities concerning both qualitative and quantitative aspects.     

Probing structural stability of double-walled carbon nanotubes at high non-hydrostatic pressure by Raman spectroscopy

Theoretical calculations predict that the collapse pressure for double-walled carbon nanotubes (DWCNTs) is proportional to 1/R ³, where R is the effective or average radius of a DWCNT. In order to address the problem of CNT stability at high pressure and stress, we performed a resonance Raman study of DWCNTs dispersed in sodium cholate using 532 and 633 nm laser excitation. Raman spectra of the recovered samples show minor versus irreversible changes with increasing I _D/I _G ratio after exposure to high non-hydrostatic pressure of 23 and 35 GPa, respectively. The system exhibits nearly 70% pressure hysteresis in radial breathing vibrational mode signals recovery on pressure release which is twice that predicted by theory.     

Continuous‐wave vertically emitting photonic crystal terahertz laser

Compact semiconductor light sources with high performance continuous‐wave (CW) and single mode operation are highly demanded for many applications in the terahertz (THz) frequency range. Distributed feedback (DFB) and photonic crystal (PhC) quantum cascade (QC) lasers are amongst the leading candidates in this field. Absorbing boundary condition is a commonly used method to control the optical performance of a laser in double‐metal confinement. However, this approach increases the total loss in the device and results in a large threshold current density, limiting the CW maximum output power and operating temperature. In this letter, a robust surface emitting continuous‐wave terahertz QC laser is realized in a two‐dimensional PhC structure by a second order Bragg grating extractor that simultaneously provides the boundary condition necessary for mode selection. This results in a 3.12 THz single mode CW operation with a 3 mW output power and a maximum operation temperature (T_max) of 100 K. Also, a highly collimated far‐field pattern is demonstrated, which is an important step towards real world applications.