Adaptive executions of hyperbolic block-structured AMR applications on GPU systems

A block-structured adaptive mesh refinement (AMR) technique has been used to obtain numerical solutions for many scientific applications. Some block-structured AMR approaches have focused on forming patches of non-uniform sizes where the size of a patch can be tuned to the geometry of a region of interest. In this paper, we develop strategies for adaptive execution of block-structured AMR applications on GPUs, for hyperbolic directionally split solvers. While effective hybrid execution strategies exist for applications with uniform patches, our work considers efficient execution of non-uniform patches with different workloads. Our techniques include bin-packing work units to load balance GPU computations, adaptive asynchronism between CPU and GPU executions using a knapsack formulation, and scheduling communications for multi-GPU executions. Our experiments with synthetic and real data, for single-GPU and multi-GPU executions, on Tesla S1070 and Fermi C2070 clusters, show that our strategies result in up to a 3.23 speedup in performance over existing strategies.

Simultaneous perturbation methods for adaptive labor staffing in service systems

We consider the problem of optimizing the workforce of a service system. Adapting the staffing levels in such systems is non-trivial due to large variations in workload and the large number of system parameters do not allow for a brute force search. Further, because these parameters change on a weekly basis, the optimization should not take longer than a few hours. Our aim is to find the optimum staffing levels from a discrete high-dimensional parameter set, that minimizes the long run average of the single-stage cost function, while adhering to the constraints relating to queue stability and service-level agreement (SLA) compliance. The single-stage cost function balances the conflicting objectives of utilizing workers better and attaining the target SLAs. We formulate this problem as a constrained parameterized Markov cost process parameterized by the (discrete) staffing levels. We propose novel simultaneous perturbation stochastic approximation (SPSA)-based algorithms for solving the above problem. The algorithms include both first-order as well as second-order methods and incorporate SPSA-based gradient/Hessian estimates for primal descent, while performing dual ascent for the Lagrange multipliers. Both algorithms are online and update the staffing levels in an incremental fashion. Further, they involve a certain generalized smooth projection operator, which is essential to project the continuous-valued worker parameter tuned by our algorithms onto the discrete set. The smoothness is necessary to ensure that the underlying transition dynamics of the constrained Markov cost process is itself smooth (as a function of the continuous-valued parameter): a critical requirement to prove the convergence of both algorithms. We validate our algorithms via performance simulations based on data from five real-life service systems. For the sake of comparison, we also implement a scatter search based algorithm using state-of-the-art optimization tool-kit OptQuest. From the experiments, we observe that both our algorithms converge empirically and consistently outperform OptQuest in most of the settings considered. This finding coupled with the computational advantage of our algorithms make them amenable for adaptive labor staffing in real-life service systems.

Timing Recovery Algorithms and Architectures for 2-D Magnetic Recording Systems

We investigate the problem of timing recovery for 2-D magnetic recording (TDMR) channels. We develop a timing error model for TDMR channel considering the phase and frequency offsets with noise. We propose a 2-D data-aided phase-locked loop (PLL) architecture for tracking variations in the position and movement of the read head in the down-track and cross-track directions and analyze the convergence of the algorithm under non-separable timing errors. We further develop a 2-D interpolation-based timing recovery scheme that works in conjunction with the 2-D PLL. We quantify the efficiency of our proposed algorithms by simulations over a 2-D magnetic recording channel with timing errors.

Throughput Analysis of Best-m Feedback in OFDM Systems with Uniformly Correlated Subchannels

Practical orthogonal frequency division multiplexing (OFDM) systems, such as Long Term Evolution (LTE), exploit multi-user diversity using very limited feedback. The best-m feedback scheme is one such limited feedback scheme, in which users report only the gains of their m best subchannels (SCs) and their indices. While the scheme has been extensively studied and adopted in standards such as LTE, an analysis of its throughput for the practically important case in which the SCs are correlated has received less attention. We derive new closed-form expressions for the throughput when the SC gains of a user are uniformly correlated. We analyze the performance of the greedy but unfair frequency-domain scheduler and the fair round-robin scheduler for the general case in which the users see statistically non-identical SCs. An asymptotic analysis is then developed to gain further insights. The analysis and extensive numerical results bring out how correlation reduces throughput.

Autonomous Star Camera Calibration and Spacecraft Attitude Determination

This paper presents two methods of star camera calibration to determine camera calibrating parameters (like principal point, focal length etc) along with lens distortions (radial and decentering). First method works autonomously utilizing star coordinates in three consecutive image frames thus independent of star identification or biased attitude information. The parameters obtained in autonomous self-calibration technique helps to identify the imaged stars with the cataloged stars. Least Square based second method utilizes inertial star coordinates to determine satellite attitude and star camera parameters with lens radial distortion, both independent of each other. Camera parameters determined by the second method are more accurate than the first method of camera self calibration. Moreover, unlike most of the attitude determination algorithms where attitude of the satellite depend on the camera calibrating parameters, the second method has the advantage of computing spacecraft attitude independent of camera calibrating parameters except lens distortions (radial). Finally Kalman filter based sequential estimation scheme is employed to filter out the noise of the LS based estimation.

Algorithm for Autonomous Reorganization of Mobile Wireless Camera Sensor Networks to Improve Coverage

In this paper, sensing coverage by wireless camera-embedded sensor networks (WCSNs), a class of directional sensors is studied. The proposed work facilitates the autonomous tuning of orientation parameters and displacement of camera-sensor nodes in the bounded field of interest (FoI), where the network coverage in terms of every point in the FoI is important. The proposed work is first of its kind to study the problem of maximizing coverage of randomly deployed mobile WCSNs which exploits their mobility. We propose an algorithm uncovered region exploration algorithm (UREA-CS) that can be executed in centralized and distributed modes. Further, the work is extended for two special scenarios: 1) to suit autonomous combing operations after initial random WCSN deployments and 2) to improve the network coverage with occlusions in the FoI. The extensive simulation results show that the performance of UREA-CS is consistent, robust, and versatile to achieve maximum coverage, both in centralized and distributed modes. The centralized and distributed modes are further analyzed with respect to the computational and communicational overheads.

Synthesis, and crystal and electronic structure of sodium metal phosphate for use as a hybrid capacitor in non-aqueous electrolyte

Energy storage devices based on sodium have been considered as an alternative to traditional lithium based systems because of the natural abundance, cost effectiveness and low environmental impact of sodium. Their synthesis, and crystal and electronic properties have been discussed, because of the importance of electronic conductivity in supercapacitors for high rate applications. The density of states of a mixed sodium transition metal phosphate (maricite, NaMn1/3Co1/3Ni1/3PO4) has been determined with the ab initio generalized gradient approximation (GGA)+Hubbard term (U) method. The computed results for the mixed maricite are compared with the band gap of the parent NaFePO4 and the electrochemical experimental results are in good agreement. A mixed sodium transition metal phosphate served as an active electrode material for a hybrid supercapacitor. The hybrid device (maricite versus carbon) in a nonaqueous electrolyte shows redox peaks in the cyclic voltammograms and asymmetric profiles in the charge-discharge curves while exhibiting a specific capacitance of 40 F g(-1) and these processes are found to be quasi-reversible. After long term cycling, the device exhibits excellent capacity retention (95%) and coulombic efficiency (92%). The presence of carbon and the nanocomposite morphology, identified through X-ray photoelectron spectroscopy (XPS) and transmission electron microscopy (TEM) studies, ensures the high rate capability while offering possibilities to develop new cathode materials for sodium hybrid devices.

Identification of nonlinear dynamic systems: Time-frequency filtering and skeleton curves

The nonlinear behavior varying with the instantaneous response was analyzed through the joint time-frequency analysis method for a class of S. D. O. F nonlinear system. A masking operator an definite regions is defined and two theorems are presented. Based on these, the nonlinear system is modeled with a special time-varying linear one, called the generalized skeleton linear system (GSLS). The frequency skeleton curve and the damping skeleton curve are defined to describe the main feature of the non-linearity as well. Moreover, an identification method is proposed through the skeleton curves and the time-frequency filtering technique.

Oscillatory instability of Rayleigh-Marangoni-Bénard convection in two-layer liquid systems

The oscillatory behaviour of the Rayleigh-Marangoni-Bénard convective instability (R-M-B instability) regarding two combinations of two-layer fluid systems has been investigated theoretically and numerically. For the two-layer system of Silicone oil (10cSt) over Fluorinert (FC70), both linear instability analysis and 2D numerical simulation show that the instability of the system depends strongly on the depth ratio Hr = H1/H2 of the two-layer liquid. The oscillatory regime at the onset of R-M-B convection enlarges with reducing Γ = Ra/Ma values. In the two-layer system of Silicone oil (2cSt) over water, it loses its stability and onsets to steady convection at first, then the steady convection bifurcates to oscillatory convection with increasing Rayleigh number Ra. This behaviour was found through numerical simulation above the onset of steady convection in the case of r = 2.9, ε=(Ra-Ruc)/Rac = 1.0, and Hr = 0.5. Our findings are different from the previous study of the Rayleigh-Benard instability and show the strong effects of the thermocapillary force at the interface on the time-dependent oscillations at or after the onset of convection. We propose a secondary oscillatory instability mechanism to explain the experimental observation of Degen et al. [Phys. Rev. E, 57 (1998), 6647-6659].

Fixed-lag blind equalization and sequence estimation in digital communications systems using sequential importance sampling

We present methods for fixed-lag smoothing using Sequential Importance sampling (SIS) on a discrete non-linear, non-Gaussian state space system with unknown parameters. Our particular application is in the field of digital communication systems. Each input data point is taken from a finite set of symbols. We represent transmission media as a fixed filter with a finite impulse response (FIR), hence a discrete state-space system is formed. Conventional Markov chain Monte Carlo (MCMC) techniques such as the Gibbs sampler are unsuitable for this task because they can only perform processing on a batch of data. Data arrives sequentially, so it would seem sensible to process it in this way. In addition, many communication systems are interactive, so there is a maximum level of latency that can be tolerated before a symbol is decoded. We will demonstrate this method by simulation and compare its performance to existing techniques.

Low-Dimensional Dynamical Systems for a Wake-Type Shear Flow with Vortex Dislocations

Low-dimensional systems are constructed to investigate dynamics of vortex dislocations in a wake-type shear flow. High-resolution direct numerical simulations are employed to obtain flow snapshots from which the most energetic modes are extracted using proper orthogonal decomposition (POD). The first 10 modes are classified into two groups. One represents the general characteristics of two-dimensional wake-type shear flow, and the other is related to the three-dimensional properties or non-uniform characteristics along the span. Vortex dislocations are generated by these two kinds of coherent structures. The results from the first 20 three-dimensional POD modes show that the low- dimensional systems have captured the basic properties of the wake-type shear flow with vortex dislocation, such as two incommensurable frequencies and their beat frequency.

Micro-systems for Optical Protein-Chip

Protein-Chip as micro-assays for the determination of protein interaction, the analysis, the identification and the purification of proteins has large potential applications. The Optical Protein-Chip is able to detect the multi-interaction of proteins and multi-bio-activities of molecules directly and simultaneously with no labeling. The chip is a small matrix on solid substrate containing multi-micro-area prepared by microfabrication with photolithography or soft lithography for surface patterning, and processed with surface modification which includes the physical, chemical, and bio-chemical modifications, etc. The ligand immobilization, such as protein immobilization, especially the oriented immobilization with low steric hindrance and high bio-specific binding activity between ligand and receptor is used to form a sensing surface. Each area of the pattern is corresponding to only one bioactivity. The interval between the areas is non-bioactive and optically extinctive. The affinity between proteins is used to realize non-labeling microassays for the determination of protein identification and protein interaction. The sampling of the chip is non-disturbing, performed with imaging ellipsometry and image processing on a database of proteins.

First-passage failure of quasi-integrable Hamiltonian systems

The first-passage failure of quasi-integrable Hamiltonian si-stems (multidegree-of-freedom integrable Hamiltonian systems subject to light dampings and weakly random excitations) is investigated. The motion equations of such a system are first reduced to a set of averaged Ito stochastic differential equations by using the stochastic averaging method for quasi-integrable Hamiltonian systems. Then, a backward Kolmogorov equation governing the conditional reliability function and a set of generalized Pontryagin equations governing the conditional moments of first-passage time are established. Finally, the conditional reliability function, and the conditional probability density and moments of first-passage time are obtained by solving these equations with suitable initial and boundary conditions. Two examples are given to illustrate the proposed procedure and the results from digital simulation are obtained to verify the effectiveness of the procedure.

On dispersion-controlled principles for non-oscillatory shock-capturing schemes

The role of dispersions in the numerical solutions of hydrodynamic equation systems has been realized for long time. It is only during the last two decades that extensive studies on the dispersion-controlled dissipative (DCD) schemes were reported. The studies have demonstrated that this kind of the schemes is distinct from conventional dissipation-based schemes in which the dispersion term of the modified equation is not considered in scheme construction to avoid nonphysical oscillation occurring in shock wave simulations. The principle of the dispersion controlled aims at removing nonphysical oscillations by making use of dispersion characteristics instead of adding artificial viscosity to dissipate the oscillation as the conventional schemes do. Research progresses on the dispersion controlled principles are reviewed in this paper, including the exploration of the role of dispersions in numerical simulations, the development of the dispersion-controlled principles, efforts devoted to high-order dispersion-controlled dissipative schemes, the extension to both the finite volume and the finite element methods, scheme verification and solution validation, and comments on several aspects of the schemes from author's viewpoint.

Optimal bounded control of first-passage failure of strongly non-linear oscillators under combined harmonic and white-noise excitations

A procedure for designing the optimal bounded control of strongly non-linear oscillators under combined harmonic and white-noise excitations for minimizing their first-passage failure is proposed. First, a stochastic averaging method for strongly non-linear oscillators under combined harmonic and white-noise excitations using generalized harmonic functions is introduced. Then, the dynamical programming equations and their boundary and final time conditions for the control problems of maximizing reliability and of maximizing mean first-passage time are formulated from the averaged Ito equations by using the dynamical programming principle. The optimal control law is derived from the dynamical programming equations and control constraint. Finally, the conditional reliability function, the conditional probability density and mean of the first-passage time of the optimally controlled system are obtained from solving the backward Kolmogorov equation and Pontryagin equation. An example is given to illustrate the proposed procedure and the results obtained are verified by using those from digital simulation. (C) 2003 Elsevier Ltd. All rights reserved.