Error, Confidence, and Cost in Engineering Computations

The questions that one should answer in engineering computations - deterministic, probabilistic/randomized, as well as heuristic - are (i) how good the computed results/outputs are and (ii) how much the cost in terms of amount of computation and the amount of storage utilized in getting the outputs is. The absolutely errorfree quantities as well as the completely errorless computations done in a natural process can never be captured by any means that we have at our disposal. While the computations including the input real quantities in nature/natural processes are exact, all the computations that we do using a digital computer or are carried out in an embedded form are never exact. The input data for such computations are also never exact because any measuring instrument has inherent error of a fixed order associated with it and this error, as a matter of hypothesis and not as a matter of assumption, is not less than 0.005 per cent. Here by error we imply relative error bounds. The fact that exact error is never known under any circumstances and any context implies that the term error is nothing but error-bounds. Further, in engineering computations, it is the relative error or, equivalently, the relative error-bounds (and not the absolute error) which is supremely important in providing us the information regarding the quality of the results/outputs. Another important fact is that inconsistency and/or near-consistency in nature, i.e., in problems created from nature is completely nonexistent while in our modelling of the natural problems we may introduce inconsistency or near-inconsistency due to human error or due to inherent non-removable error associated with any measuring device or due to assumptions introduced to make the problem solvable or more easily solvable in practice. Thus if we discover any inconsistency or possibly any near-inconsistency in a mathematical model, it is certainly due to any or all of the three foregoing factors. We do, however, go ahead to solve such inconsistent/near-consistent problems and do get results that could be useful in real-world situations. The talk considers several deterministic, probabilistic, and heuristic algorithms in numerical optimisation, other numerical and statistical computations, and in PAC (probably approximately correct) learning models. It highlights the quality of the results/outputs through specifying relative error-bounds along with the associated confidence level, and the cost, viz., amount of computations and that of storage through complexity. It points out the limitation in error-free computations (wherever possible, i.e., where the number of arithmetic operations is finite and is known a priori) as well as in the usage of interval arithmetic. Further, the interdependence among the error, the confidence, and the cost is discussed.

Decoupling of diffusion from viscosity: Difference scenario for translational and rotational motions

Recent experiments have indicated a dramatically different viscosity dependence of the translational and the rotational diffusion coefficients in a supercooled liquid as the glass transition temperature is approached from above. While the translational motion seems to be decoupled from the rising viscosity (eta), the rotational motion seems to remain firmly coupled to eta. In order to understand the microscopic origin of this behavior, we have carried nut detailed theoretical calculations of both the quantities by using a self-consistent mode-coupling theory (MCT). it is found that when the size of the solute is same as that of the solvent molecules, the conventional MCT fails to predict the observed decoupling. The solvent inhomogeneity is found to play a decisive role in determining the decoupling. The difference in the viscosity dependence between rotation and translational diffusion coefficient is discussed.

Self-diffusion of colloids at freezing

This is an introduction to the theory of interacting Brownian particles, as applied to charge-stabilised colloidal suspensions near their equilibrium liquid-solid transition. The density functional approach to the statics of the transition is reviewed briefly, and the generalised Langevin equation method for the dynamics presented in detail. Work with A.V. Indrani [1] on a self-consistent approach for calculating the excess single-particle friction is presented, which explains the observed [2] ''universal'' suppression of self-diffusion at freezing as a consequence of the universal structure-factor height at this transition. Criticisms, open questions, and challenges to theory are discussed.

Shear-enhanced diffusion in charge-stabilised colloids

The generalised Langevin equation method for the dynamics of interacting colloids presented in my previous lecture is extended here to the case of a sheared suspension. A calculation of shear-dependent diffusivities using these methods is found to account for puzzling observations in experiments and simulations. The limitations of the method are discussed, and important unresolved questions presented. This lecture summarises work done in collaboration with A.V. Indrani [1].

Monsoon Regimes and Processes in CCSM4. Part I: The Asian-Australian Monsoon

The simulation characteristics of the Asian-Australian monsoon are documented for the Community Climate System Model, version 4 (CCSM4). This is the first part of a two part series examining monsoon regimes in the global tropics in the CCSM4. Comparisons are made to an Atmospheric Model Intercomparison Project (AMIP) simulation of the atmospheric component in CCSM4 Community Atmosphere Model, version 4, (CAM4)] to deduce differences in the monsoon simulations run with observed sea surface temperatures (SSTs) and with ocean-atmosphere coupling. These simulations are also compared to a previous version of the model (CCSM3) to evaluate progress. In general, monsoon rainfall is too heavy in the uncoupled AMIP run with CAM4, and monsoon rainfall amounts are generally better simulated with ocean coupling in CCSM4. Most aspects of the Asian-Australian monsoon simulations are improved in CCSM4 compared to CCSM3. There is a reduction of the systematic error of rainfall over the tropical Indian Ocean for the South Asian monsoon, and well-simulated connections between SSTs in the Bay of Bengal and regional South Asian monsoon precipitation. The pattern of rainfall in the Australian monsoon is closer to observations in part because of contributions from the improvements of the Indonesian Throughflow and diapycnal diffusion in CCSM4. Intraseasonal variability of the Asian-Australian monsoon is much improved in CCSM4 compared to CCSM3 both in terms of eastward and northward propagation characteristics, though it is still somewhat weaker than observed. An improved simulation of El Nino in CCSM4 contributes to more realistic connections between the Asian-Australian monsoon and El Nino-Southern Oscillation (ENSO), though there is considerable decadal and century time scale variability of the strength of the monsoon-ENSO connection.

Online Computation of Hysteresis Boundary for Constant Switching Frequency Current-Error Space-Vector-Based Hysteresis Controller for VSI-Fed IM Drives

This paper proposes a current-error space-vector-based hysteresis controller with online computation of boundary for two-level inverter-fed induction motor (IM) drives. The proposed hysteresis controller has got all advantages of conventional current-error space-vector-based hysteresis controllers like quick transient response, simplicity, adjacent voltage vector switching, etc. Major advantage of the proposed controller-based voltage-source-inverters-fed drive is that phase voltage frequency spectrum produced is exactly similar to that of a constant switching frequency space-vector pulsewidth modulated (SVPWM) inverter. In this proposed hysteresis controller, stator voltages along alpha- and beta-axes are estimated during zero and active voltage vector periods using current errors along alpha- and beta-axes and steady-state model of IM. Online computation of hysteresis boundary is carried out using estimated stator voltages in the proposed hysteresis controller. The proposed scheme is simple and capable of taking inverter upto six-step-mode operation, if demanded by drive system. The proposed hysteresis-controller-based inverter-fed drive scheme is experimentally verified. The steady state and transient performance of the proposed scheme is extensively tested. The experimental results are giving constant frequency spectrum for phase voltage similar to that of constant frequency SVPWM inverter-fed drive.

Diffusion ordered spectroscopy for resolution of double bonded cis, trans-isomers

NMR spectroscopic separation of double bonded cis- and trans-isomers, that have different molecular shapes but identical mass have been carried out using Diffusion Ordered Spectroscopy (DOSY). The mixtures of fumaric acid and maleic acid, that have similar hydrodynamic radii, have resolved been on the basis of their diffusion coefficients arising due to their different tendencies to associate with micelles or reverse micelles. Sodium dodecyl sulfate (SDS) and Dioctyl sulfosuccinate sodium salt (AOT) have been used as the media to mimic the chromatographic conditions, modify the average mobility and to achieve differential diffusion rates. The best separation of the components has been achieved by Dioctyl sulfosuccinate sodium salt (AOT) in D2O solution. (C) 2012 Elsevier B.V. All rights reserved.

Linearization and Reconstruction of Non-linear Diffuse Optical Tomographic Image

Diffuse optical tomography (DOT) is one of the ways to probe highly scattering media such as tissue using low-energy near infra-red light (NIR) to reconstruct a map of the optical property distribution. The interaction of the photons in biological tissue is a non-linear process and the phton transport through the tissue is modelled using diffusion theory. The inversion problem is often solved through iterative methods based on nonlinear optimization for the minimization of a data-model misfit function. The solution of the non-linear problem can be improved by modeling and optimizing the cost functional. The cost functional is f(x) = x(T)Ax - b(T)x + c and after minimization, the cost functional reduces to Ax = b. The spatial distribution of optical parameter can be obtained by solving the above equation iteratively for x. As the problem is non-linear, ill-posed and ill-conditioned, there will be an error or correction term for x at each iteration. A linearization strategy is proposed for the solution of the nonlinear ill-posed inverse problem by linear combination of system matrix and error in solution. By propagating the error (e) information (obtained from previous iteration) to the minimization function f(x), we can rewrite the minimization function as f(x; e) = (x + e)(T) A(x + e) - b(T)(x + e) + c. The revised cost functional is f(x; e) = f(x) + e(T)Ae. The self guided spatial weighted prior (e(T)Ae) error (e, error in estimating x) information along the principal nodes facilitates a well resolved dominant solution over the region of interest. The local minimization reduces the spreading of inclusion and removes the side lobes, thereby improving the contrast, localization and resolution of reconstructed image which has not been possible with conventional linear and regularization algorithm.

Robust solver based on modified particle swarm optimization for improved solution of diffusion transport through containment facilities

Accurate estimation of mass transport parameters is necessary for overall design and evaluation processes of the waste disposal facilities. The mass transport parameters, such as effective diffusion coefficient, retardation factor and diffusion accessible porosity, are estimated from observed diffusion data by inverse analysis. Recently, particle swarm optimization (PSO) algorithm has been used to develop inverse model for estimating these parameters that alleviated existing limitations in the inverse analysis. However, PSO solver yields different solutions in successive runs because of the stochastic nature of the algorithm and also because of the presence of multiple optimum solutions. Thus the estimated mean solution from independent runs is significantly different from the best solution. In this paper, two variants of the PSO algorithms are proposed to improve the performance of the inverse analysis. The proposed algorithms use perturbation equation for the gbest particle to gain information around gbest region on the search space and catfish particles in alternative iterations to improve exploration capabilities. Performance comparison of developed solvers on synthetic test data for two different diffusion problems reveals that one of the proposed solvers, CPPSO, significantly improves overall performance with improved best, worst and mean fitness values. The developed solver is further used to estimate transport parameters from 12 sets of experimentally observed diffusion data obtained from three diffusion problems and compared with published values from the literature. The proposed solver is quick, simple and robust on different diffusion problems. (C) 2012 Elsevier Ltd. All rights reserved.

Diffusion mechanism in XSi2 and X5Si3 (X= Nb, Mo, V) phases

In view of the importance of the suicides in the high temperature applications, the diffusion behaviour is compared in different systems for two types of silicides, XSi2 and X5Si3 (X=Nb, Mo, V). Atomic mechanism of diffusion and defects present in the structure are discussed. In both the phases, Si has faster diffusion rate than the metal species. This is expected from the nearest neighbour (NN) bonds present in the XSi2 phase but rather unusual in the X5Si3 phase. Relative mobilities of the species calculated indicate the presence of high concentration of Si antisites. Moreover, the concentration of the defects is different in different systems to find different diffusion rates.

Diffusion mechanism in the mu phase in Nb-X (X = Ni, Co, Fe) systems

Nb is one of the common refractory elements added in Ni, Co and Fe based superalloys. This lead to the formation of brittle topological close packed (tcp) mu phase, which is deleterious to the structure. It mainly grows by interdiffusion and in the present article, the interdiffusion process in different Nb-X (X=Ni, Co, Fe) systems is discussed. The activation energy for interdiffusion is lower in the Co-Nb system (173 kJ/mol) than Fe-Nb system (233 kJ/mol), which is again lower than the value found in the Ni-Nb system (319.7 kJ/mol). The mole fraction of Nb in this phase is less than Fe or Co at stoichiometric compositions in the Nb-Fe (that is Fe7Nb6) and Nb-Co (that is Co7Nb6) systems. On the other hand, the mole fraction of Nb is higher than Ni in the same phase (Ni6Nb2) in Ni-Nb system. However, in all the phases, Nb has lower diffusion rate. Possible diffusion mechanism in this phase is discussed with respect to the crystal structure.

A versatile resolving agent for diffusion edited separation of enantiomers, complex mixtures and constitutional isomers

The study is the first report of the utilization of a crown ether as a new and versatile resolving agent for the diffusion edited separation of enantiomers, complex mixtures and constitutional isomers. As a consequence of different binding affinities of enantiomers of a chiral molecule and individual components of the complex mixtures with the crown ether, the molecules diffuse at different rates. The enhanced separation achieved due to matrix assisted diffusion permitted their separation in the diffusion dimension. The generality and wide utility of the new resolving agent and the methodology are demonstrated on diverse examples, such as an organic chiral molecule, constitutional isomers and complex mixture of molecules possessing different functional groups that possess nearly identical molecular weights.

Derating Based Hardware Optimizations in Soft Error Tolerant Designs

Ensuring reliable operation over an extended period of time is one of the biggest challenges facing present day electronic systems. The increased vulnerability of the components to atmospheric particle strikes poses a big threat in attaining the reliability required for various mission critical applications. Various soft error mitigation methodologies exist to address this reliability challenge. A general solution to this problem is to arrive at a soft error mitigation methodology with an acceptable implementation overhead and error tolerance level. This implementation overhead can then be reduced by taking advantage of various derating effects like logical derating, electrical derating and timing window derating, and/or making use of application redundancy, e. g. redundancy in firmware/software executing on the so designed robust hardware. In this paper, we analyze the impact of various derating factors and show how they can be profitably employed to reduce the hardware overhead to implement a given level of soft error robustness. This analysis is performed on a set of benchmark circuits using the delayed capture methodology. Experimental results show upto 23% reduction in the hardware overhead when considering individual and combined derating factors.

Stochastic Simulations Suggest that HIV-1 Survives Close to Its Error Threshold

The use of mutagenic drugs to drive HIV-1 past its error threshold presents a novel intervention strategy, as suggested by the quasispecies theory, that may be less susceptible to failure via viral mutation-induced emergence of drug resistance than current strategies. The error threshold of HIV-1, mu(c), however, is not known. Application of the quasispecies theory to determine mu(c) poses significant challenges: Whereas the quasispecies theory considers the asexual reproduction of an infinitely large population of haploid individuals, HIV-1 is diploid, undergoes recombination, and is estimated to have a small effective population size in vivo. We performed population genetics-based stochastic simulations of the within-host evolution of HIV-1 and estimated the structure of the HIV-1 quasispecies and mu(c). We found that with small mutation rates, the quasispecies was dominated by genomes with few mutations. Upon increasing the mutation rate, a sharp error catastrophe occurred where the quasispecies became delocalized in sequence space. Using parameter values that quantitatively captured data of viral diversification in HIV-1 patients, we estimated mu(c) to be 7 x 10(-5) -1 x 10(-4) substitutions/site/replication, similar to 2-6 fold higher than the natural mutation rate of HIV-1, suggesting that HIV-1 survives close to its error threshold and may be readily susceptible to mutagenic drugs. The latter estimate was weakly dependent on the within-host effective population size of HIV-1. With large population sizes and in the absence of recombination, our simulations converged to the quasispecies theory, bridging the gap between quasispecies theory and population genetics-based approaches to describing HIV-1 evolution. Further, mu(c) increased with the recombination rate, rendering HIV-1 less susceptible to error catastrophe, thus elucidating an added benefit of recombination to HIV-1. Our estimate of mu(c) may serve as a quantitative guideline for the use of mutagenic drugs against HIV-1.