We present a novel class of dynamic neural networks that is capable of learning, in an unsupervised manner, attractors that correspond to generalities in a data set. Upon presentation of a test stimulus, the networks follow a sequence of attractors that correspond to subsets of increasing size or generality in the original data set. The networks, inspired by those of the insect antennal lobe, build upon a modified Hopfield network in which nodes are periodically suppressed, global inhibition is gradually strengthened, and the weight of input neurons is gradually decreased relative to recurrent connections. This allows the networks to converge on a Hopfield network's equilibrium within each suppression cycle, and to switch between attractors in between cycles. The fast mutually reinforcing excitatory connections that dominate dynamics within cycles ensures the robust error-tolerant behavior that characterizes Hopfield networks. The cyclic inhibition releases the network from what would otherwise be stable equilibriums or attractors. Increasing global inhibition and decreasing dependence on the input leads successive attractors to differ, and to display increasing generality. As the network is faced with stronger inhibition, only neurons connected with stronger mutually excitatory connections will remain on; successive attractors will consist of sets of neurons that are more strongly correlated, and will tend to select increasingly generic characteristics of the data. Using artificial data, we were able to identify configurations of the network that appeared to produce a sequence of increasingly general results. The next logical steps are to apply these networks to suitable real-world data that can be characterized by a hierarchy of increasing generality and observe the network's performance. This report describes the work, data, and results, the current understanding of the results, and how the work could be continued. The code, data, and preliminary results are included and are available as an archive.
We developed an Augmented Musculature Device (AMD) that assists the movements of its wearer. It has direct application to aiding military and law enforcement personnel, the neurologically impaired, or those requiring any type of cybernetic assistance. The AMD consists of a collection of artificial muscles, each individually actuated, strategically placed along the surface of the human body. The actuators employed by the AMD are known as 'air muscles' and operate pneumatically. They are commercially available from several vendors and are relatively inexpensive. They have a remarkably high force-to-weight ratio--as high as 400:1 (as compared with 16:1 typical of DC motors). They are flexible and elastic, even when powered, making them ideal for interaction with humans.
The challenge of modeling the organization and function of biological membranes on a solid support has received considerable attention in recent years, primarily driven by potential applications in biosensor design. Affinity-based biosensors show great promise for extremely sensitive detection of BW agents and toxins. Receptor molecules have been successfully incorporated into phospholipid bilayers supported on sensing platforms. However, a collective body of data detailing a mechanistic understanding of membrane processes involved in receptor-substrate interactions and the competition between localized perturbations and delocalized responses resulting in reorganization of transmembrane protein structure, has yet to be produced. This report describes a systematic procedure to develop detailed correlation between (recognition-induced) protein restructuring and function of a ligand gated ion channel by combining single molecule fluorescence spectroscopy and single channel current recordings. This document is divided into three sections: (1) reported are the thermodynamics and diffusion properties of gramicidin using single molecule fluorescence imaging and (2) preliminary work on the 5HT{sub 3} serotonin receptor. Thirdly, we describe the design and fabrication of a miniaturized platform using the concepts of these two technologies (spectroscopic and single channel electrochemical techniques) for single molecule analysis, with a longer term goal of using the physical and electronic changes caused by a specific molecular recognition event as a transduction pathway in affinity based biosensors for biotoxin detection.
This white paper represents a summary of work intended to lay the foundation for development of a climatological/agent model of climate-induced conflict. The paper combines several loosely-coupled efforts and is the final report for a four-month late-start Laboratory Directed Research and Development (LDRD) project funded by the Advanced Concepts Group (ACG). The project involved contributions by many participants having diverse areas of expertise, with the common goal of learning how to tie together the physical and human causes and consequences of climate change. We performed a review of relevant literature on conflict arising from environmental scarcity. Rather than simply reviewing the previous work, we actively collected data from the referenced sources, reproduced some of the work, and explored alternative models. We used the unfolding crisis in Darfur (western Sudan) as a case study of conflict related to or triggered by climate change, and as an exercise for developing a preliminary concept map. We also outlined a plan for implementing agents in a climate model and defined a logical progression toward the ultimate goal of running both types of models simultaneously in a two-way feedback mode, where the behavior of agents influences the climate and climate change affects the agents. Finally, we offer some ''lessons learned'' in attempting to keep a diverse and geographically dispersed group working together by using Web-based collaborative tools.
The Common Geometry Module (CGM) is a code library which provides geometry functionality used for mesh generation and other applications. This functionality includes that commonly found in solid modeling engines, like geometry creation, query and modification; CGM also includes capabilities not commonly found in solid modeling engines, like geometry decomposition tools and support for shared material interfaces. CGM is built upon the ACIS solid modeling engine, but also includes geometry capability developed beside and on top of ACIS. CGM can be used as-is to provide geometry functionality for codes needing this capability. However, CGM can also be extended using derived classes in C++, allowing the geometric model to serve as the basis for other applications, for example mesh generation. CGM is supported on Sun Solaris, SGI, HP, IBM, DEC, Linux and Windows NT platforms. CGM also includes support for loading ACIS models on parallel computers, using MPI-based communication. Future plans for CGM are to port it to different solid modeling engines, including Pro/Engineer or SolidWorks. CGM is being released into the public domain under an LGPL license; the ACIS-based engine is available to ACIS licensees on request.
Thermal actuators have proven to be a robust actuation method in surface-micromachined MEMS processes. Their higher output force and lower input voltage make them an attractive alternative to more traditional electrostatic actuation methods. A predictive model of thermal actuator behavior has been developed and validated that can be used as a design tool to customize the performance of an actuator to a specific application. This tool has also been used to better understand thermal actuator reliability by comparing the maximum actuator temperature to the measured lifetime. Modeling thermal actuator behavior requires the use of two sequentially coupled models, the first to predict the temperature increase of the actuator due to the applied current and the second to model the mechanical response of the structure due to the increase in temperature. These two models have been developed using Matlab for the thermal response and ANSYS for the structural response. Both models have been shown to agree well with experimental data. In a parallel effort, the reliability and failure mechanisms of thermal actuators have been studied. Their response to electrical overstress and electrostatic discharge has been measured and a study has been performed to determine actuator lifetime at various temperatures and operating conditions. The results from this study have been used to determine a maximum reliable operating temperature that, when used in conjunction with the predictive model, enables us to design in reliability and customize the performance of an actuator at the design stage.
The solution of the governing steady transport equations for momentum, heat and mass transfer in fluids undergoing non-equilibrium chemical reactions can be extremely challenging. The difficulties arise from both the complexity of the nonlinear solution behavior as well as the nonlinear, coupled, non-symmetric nature of the system of algebraic equations that results from spatial discretization of the PDEs. In this paper, we briefly review progress on developing a stabilized finite element ( FE) capability for numerical solution of these challenging problems. The discussion considers the stabilized FE formulation for the low Mach number Navier-Stokes equations with heat and mass transport with non-equilibrium chemical reactions, and the solution methods necessary for detailed analysis of these complex systems. The solution algorithms include robust nonlinear and linear solution schemes, parameter continuation methods, and linear stability analysis techniques. Our discussion considers computational efficiency, scalability, and some implementation issues of the solution methods. Computational results are presented for a CFD benchmark problem as well as for a number of large-scale, 2D and 3D, engineering transport/reaction applications.
This study investigates algebraic multilevel domain decomposition preconditioners of the Schwarz type for solving linear systems associated with Newton-Krylov methods. The key component of the preconditioner is a coarse approximation based on algebraic multigrid ideas to approximate the global behavior of the linear system. The algebraic multilevel preconditioner is based on an aggressive coarsening graph partitioning of the non-zero block structure of the Jacobian matrix. The scalability of the preconditioner is presented as well as comparisons with a two-level Schwarz preconditioner using a geometric coarse grid operator. These comparisons are obtained on large-scale distributed-memory parallel machines for systems arising from incompressible flow and transport using a stabilized finite element formulation. The results demonstrate the influence of the smoothers and coarse level solvers for a set of 3D example problems. For preconditioners with more than one level, careful attention needs to be given to the balance of robustness and convergence rate for the smoothers and the cost of applying these methods. For properly chosen parameters, the two- and three-level preconditioners are demonstrated to be scalable to 1024 processors.
Least-squares finite-element methods for Darcy flow offer several advantages relative to the mixed-Galerkin method: the avoidance of stability conditions between finite-element spaces, the efficiency of solving symmetric and positive definite systems, and the convenience of using standard, continuous nodal elements for all variables. However, conventional C{sup o} implementations conserve mass only approximately and for this reason they have found limited acceptance in applications where locally conservative velocity fields are of primary interest. In this paper, we show that a properly formulated compatible least-squares method offers the same level of local conservation as a mixed method. The price paid for gaining favourable conservation properties is that one has to give up what is arguably the least important advantage attributed to least-squares finite-element methods: one can no longer use continuous nodal elements for all variables. As an added benefit, compatible least-squares methods inherit the best computational properties of both Galerkin and mixed-Galerkin methods and, in some cases, yield identical results, while offering the advantages of not having to deal with stability conditions and yielding positive definite discrete problems. Numerical results that illustrate our findings are provided.
Acts of terrorism could have a range of broad impacts on an economy, including changes in consumer (or demand) confidence and the ability of productive sectors to respond to changes. As a first step toward a model of terrorism-based impacts, we develop here a model of production and employment that characterizes dynamics in ways useful toward understanding how terrorism-based shocks could propagate through the economy; subsequent models will introduce the role of savings and investment into the economy. We use Aspen, a powerful economic modeling tool developed at Sandia, to demonstrate for validation purposes that a single-firm economy converges to the known monopoly equilibrium price, output, and employment levels, while multiple-firm economies converge toward the competitive equilibria typified by lower prices and higher output and employment. However, we find that competition also leads to churn by consumers seeking lower prices, making it difficult for firms to optimize with respect to wages, prices, and employment levels. Thus, competitive firms generate market ''noise'' in the steady state as they search for prices and employment levels that will maximize profits. In the context of this model, not only could terrorism depress overall consumer confidence and economic activity but terrorist acts could also cause normal short-run dynamics to be misinterpreted by consumers as a faltering economy.
The impact of 3D structure on wire array z-pinch dynamics is a topic of current interest, and has been studied by the controlled seeding of wire perturbations. First, Al wires were etched at Sandia, creating 20% radial perturbations with variable axial wavelength. Observations of magnetic bubble formation in the etched regions during experiments on the MAGPIE accelerator are discussed and compared to 3D MHD modeling. Second, thin NaF coatings of 1 mm axial extent were deposited on Al wires and fielded on the Zebra accelerator. Little or no axial transport of the NaF spectroscopic dopant was observed in spatially resolved K-shell spectra, which places constraints on particle diffusivity in dense z-pinch plasmas. Finally, technology development for seeding perturbations is discussed.
This paper builds upon previous work [Sprigg and Ehlen, 2004] by introducing a bond market into a model of production and employment. The previous paper described an economy in which households choose whether to enter the labor and product markets based on wages and prices. Firms experiment with prices and employment levels to maximize their profits. We developed agent-based simulations using Aspen, a powerful economic modeling tool developed at Sandia, to demonstrate that multiple-firm economies converge toward the competitive equilibria typified by lower prices and higher output and employment, but also suffer from market noise stemming from consumer churn. In this paper we introduce a bond market as a mechanism for household savings. We simulate an economy of continuous overlapping generations in which each household grows older in the course of the simulation and continually revises its target level of savings according to a life-cycle hypothesis. Households can seek employment, earn income, purchase goods, and contribute to savings until they reach the mandatory retirement age; upon retirement households must draw from savings in order to purchase goods. This paper demonstrates the simultaneous convergence of product, labor, and savings markets to their calculated equilibria, and simulates how a disruption to a productive sector will create cascading effects in all markets. Subsequent work will use similar models to simulate how disruptions, such as terrorist attacks, would interplay with consumer confidence to affect financial markets and the broader economy.
In the search for ''good'' parallel programming environments for Sandia's current and future parallel architectures, they revisit a long-standing open question. Can the PRAM parallel algorithms designed by theoretical computer scientists over the last two decades be implemented efficiently? This open question has co-existed with ongoing efforts in the HPC community to develop practical parallel programming models that can simultaneously provide ease of use, expressiveness, performance, and scalability. Unfortunately, no single model has met all these competing requirements. Here they propose a parallel programming environment, PRAM C, to bridge the gap between theory and practice. This is an attempt to provide an affirmative answer to the PRAM question, and to satisfy these competing practical requirements. This environment consists of a new thin runtime layer and an ANSI C extension. The C extension has two control constructs and one additional data type concept, ''shared''. This C extension should enable easy translation from PRAM algorithms to real parallel programs, much like the translation from sequential algorithms to C programs. The thin runtime layer bundles fine-grained communication requests into coarse-grained communication to be served by message-passing. Although the PRAM represents SIMD-style fine-grained parallelism, a stand-alone PRAM C environment can support both fine-grained and coarse-grained parallel programming in either a MIMD or SPMD style, interoperate with existing MPI libraries, and use existing hardware. The PRAM C model can also be integrated easily with existing models. Unlike related efforts proposing innovative hardware with the goal to realize the PRAM, ours can be a pure software solution with the purpose to provide a practical programming environment for existing parallel machines; it also has the potential to perform well on future parallel architectures.
Genetic programming (GP) has proved to be a highly versatile and useful tool for identifying relationships in data for which a more precise theoretical construct is unavailable. In this project, we use a GP search to develop trading strategies for agent based economic models. These strategies use stock prices and technical indicators, such as the moving average convergence/divergence and various exponentially weighted moving averages, to generate buy and sell signals. We analyze the effect of complexity constraints on the strategies as well as the relative performance of various indicators. We also present innovations in the classical genetic programming algorithm that appear to improve convergence for this problem. Technical strategies developed by our GP algorithm can be used to control the behavior of agents in economic simulation packages, such as ASPEN-D, adding variety to the current market fundamentals approach. The exploitation of arbitrage opportunities by technical analysts may help increase the efficiency of the simulated stock market, as it does in the real world. By improving the behavior of simulated stock markets, we can better estimate the effects of shocks to the economy due to terrorism or natural disasters.
In this paper we present an analysis of a new configuration for achieving spin stabilized magnetic levitation. In the classical configuration, the rotor spins about a vertical axis; and the spin stabilizes the lateral instability of the top in the magnetic field. In this new configuration the rotor spins about a horizontal axis; and the spin stabilizes the axial instability of the top in the magnetic field.
ML is a multigrid preconditioning package intended to solve linear systems of equations Ax = b where A is a user supplied n x n sparse matrix, b is a user supplied vector of length n and x is a vector of length n to be computed. ML should be used on large sparse linear systems arising from partial differential equation (PDE) discretizations. While technically any linear system can be considered, ML should be used on linear systems that correspond to things that work well with multigrid methods (e.g. elliptic PDEs). ML can be used as a stand-alone package or to generate preconditioners for a traditional iterative solver package (e.g. Krylov methods). We have supplied support for working with the Aztec 2.1 and AztecOO iterative package [16]. However, other solvers can be used by supplying a few functions. This document describes one specific algebraic multigrid approach: smoothed aggregation. This approach is used within several specialized multigrid methods: one for the eddy current formulation for Maxwell's equations, and a multilevel and domain decomposition method for symmetric and nonsymmetric systems of equations (like elliptic equations, or compressible and incompressible fluid dynamics problems). Other methods exist within ML but are not described in this document. Examples are given illustrating the problem definition and exercising multigrid options.
This paper presents solution verification studies applicable to a class of problems involving wave propagation, frictional contact, geometrical complexity, and localized incompressibility. The studies are in support of a validation exercise of a phenomenological screw failure model. The numerical simulations are performed using a fully explicit transient dynamics finite element code, employing both standard four-node tetrahedral and eight-node mean quadrature hexahedral elements. It is demonstrated that verifying the accuracy of the simulation involves not only consideration of the mesh discretization error, but also the effect of the hourglass control and the contact enforcement. In particular, the proper amount of hourglass control and the behavior of the contact search and enforcement algorithms depend greatly on the mesh resolution. We carry out the solution verification exercise using mesh refinement studies and describe our systematic approach to handling the complicating issues. It is shown that hourglassing and contact must both be carefully monitored as the mesh is refined, and it is often necessary to make adjustments to the hourglass and contact user input parameters to accommodate finer meshes. We introduce in this paper the hourglass energy, which is used as an 'error indicator' for the hourglass control. If the hourglass energy does not tend to zero with mesh refinement, then an hourglass control parameter is changed and the calculation is repeated.
An important challenge encountered during post-processing of finite element analyses is the visualizing of three-dimensional fields of real-valued second-order tensors. Namely, as finite element meshes become more complex and detailed, evaluation and presentation of the principal stresses becomes correspondingly problematic. In this paper, we describe techniques used to visualize simulations of perturbed in-situ stress fields associated with hypothetical salt bodies in the Gulf of Mexico. We present an adaptation of the Mohr diagram, a graphical paper and pencil method used by the material mechanics community for estimating coordinate transformations for stress tensors, as a new tensor glyph for dynamically exploring tensor variables within three-dimensional finite element models. This interactive glyph can be used as either a probe or a filter through brushing and linking.
The purpose of the present work is to increase our understanding of which properties of geomaterials most influence the penetration process with a goal of improving our predictive ability. Two primary approaches were followed: development of a realistic, constitutive model for geomaterials and designing an experimental approach to study penetration from the target's point of view. A realistic constitutive model, with parameters based on measurable properties, can be used for sensitivity analysis to determine the properties that are most important in influencing the penetration process. An immense literature exists that is devoted to the problem of predicting penetration into geomaterials or similar man-made materials such as concrete. Various formulations have been developed that use an analytic or more commonly, numerical, solution for the spherical or cylindrical cavity expansion as a sort of Green's function to establish the forces acting on a penetrator. This approach has had considerable success in modeling the behavior of penetrators, both as to path and depth of penetration. However the approach is not well adapted to the problem of understanding what is happening to the material being penetrated. Without a picture of the stress and strain state imposed on the highly deformed target material, it is not easy to determine what properties of the target are important in influencing the penetration process. We developed an experimental arrangement that allows greater control of the deformation than is possible in actual penetrator tests, yet approximates the deformation processes imposed by a penetrator. Using explosive line charges placed in a central borehole, we loaded cylindrical specimens in a manner equivalent to an increment of penetration, allowing the measurement of the associated strains and accelerations and the retrieval of specimens from the more-or-less intact cylinder. Results show clearly that the deformation zone is highly concentrated near the borehole, with almost no damage occurring beyond 1/2 a borehole diameter. This implies penetration is not strongly influenced by anything but the material within a diameter or so of the penetration. For penetrator tests, target size should not matter strongly once target diameters exceed some small multiple of the penetrator diameter. Penetration into jointed rock should not be much affected unless a discontinuity is within a similar range. Accelerations measured at several points along a radius from the borehole are consistent with highly-concentrated damage and energy absorption; At the borehole wall, accelerations were an order of magnitude higher than at 1/2 a diameter, but at the outer surface, 8 diameters away, accelerations were as expected for propagation through an elastic medium. Accelerations measured at the outer surface of the cylinders increased significantly with cure time for the concrete. As strength increased, less damage was observed near the explosively-driven borehole wall consistent with the lower energy absorption expected and observed for stronger concrete. As it is the energy absorbing properties of a target that ultimately stop a penetrator, we believe this may point the way to a more readily determined equivalent of the S number.
Biosecurity must be implemented without impeding biomedical and bioscience research. Existing security literature and regulatory requirements do not present a comprehensive approach or clear model for biosecurity, nor do they wholly recognize the operational issues within laboratory environments. To help address these issues, the concept of Biosecurity Levels should be developed. Biosecurity Levels would have increasing levels of security protections depending on the attractiveness of the pathogens to adversaries. Pathogens and toxins would be placed in a Biosecurity Level based on their security risk. Specifically, the security risk would be a function of an agent's weaponization potential and consequences of use. To demonstrate the concept, examples of security risk assessments for several human, animal, and plant pathogens will be presented. Higher security than that currently mandated by federal regulations would be applied for those very few agents that represent true weapons threats and lower levels for the remainder.
By means of coupled-cluster theory, molecular properties can be computed with an accuracy often exceeding that of experiment. The high-degree polynomial scaling of the coupled-cluster method, however, remains a major obstacle in the accurate theoretical treatment of mainstream chemical problems, despite tremendous progress in computer architectures. Although it has long been recognized that this super-linear scaling is non-physical, the development of efficient reduced-scaling algorithms for massively parallel computers has not been realized. We here present a locally correlated, reduced-scaling, massively parallel coupled-cluster algorithm. A sparse data representation for handling distributed, sparse multidimensional arrays has been implemented along with a set of generalized contraction routines capable of handling such arrays. The parallel implementation entails a coarse-grained parallelization, reducing interprocessor communication and distributing the largest data arrays but replicating as many arrays as possible without introducing memory bottlenecks. The performance of the algorithm is illustrated by several series of runs for glycine chains using a Linux cluster with an InfiniBand interconnect.
We explore stability of Random Boolean Networks as a model of biological interaction networks. We introduce surface-to-volume ratio as a measure of stability of the network. Surface is defined as the set of states within a basin of attraction that maps outside the basin by a bit-flip operation. Volume is defined as the total number of states in the basin. We report development of an object-oriented Boolean network analysis code (Attract) to investigate the structure of stable vs. unstable networks. We find two distinct types of stable networks. The first type is the nearly trivial stable network with a few basins of attraction. The second type contains many basins. We conclude that second type stable networks are extremely rare.
The efficiency of neuronal encoding in sensory and motor systems has been proposed as a first principle governing response properties within the central nervous system. We present a continuation of a theoretical study presented by Zhang and Sejnowski, where the influence of neuronal tuning properties on encoding accuracy is analyzed using information theory. When a finite stimulus space is considered, we show that the encoding accuracy improves with narrow tuning for one- and two-dimensional stimuli. For three dimensions and higher, there is an optimal tuning width.
Human behavior is a function of an iterative interaction between the stimulus environment and past experience. It is not simply a matter of the current stimulus environment activating the appropriate experience or rule from memory (e.g., if it is dark and I hear a strange noise outside, then I turn on the outside lights and investigate). Rather, it is a dynamic process that takes into account not only things one would generally do in a given situation, but things that have recently become known (e.g., there have recently been coyotes seen in the area and one is known to be rabid), as well as other immediate environmental characteristics (e.g., it is snowing outside, I know my dog is outside, I know the police are already outside, etc.). All of these factors combine to inform me of the most appropriate behavior for the situation. If it were the case that humans had a rule for every possible contingency, the amount of storage that would be required to enable us to fluidly deal with most situations we encounter would rapidly become biologically untenable. We can all deal with contingencies like the one above with fairly little effort, but if it isn't based on rules, what is it based on? The assertion of the Cognitive Systems program at Sandia for the past 5 years is that at the heart of this ability to effectively navigate the world is an ability to discriminate between different contexts (i.e., Dynamic Context Discrimination, or DCD). While this assertion in and of itself might not seem earthshaking, it is compelling that this ability and its components show up in a wide variety of paradigms across different subdisciplines in psychology. We begin by outlining, at a high functional level, the basic ideas of DCD. We then provide evidence from several different literatures and paradigms that support our assertion that DCD is a core aspect of cognitive functioning. Finally, we discuss DCD and the computational model that we have developed as an instantiation of DCD in more detail. Before commencing with our overview of DCD, we should note that DCD is not necessarily a theory in the classic sense. Rather, it is a description of cognitive functioning that seeks to unify highly similar findings across a wide variety of literatures. Further, we believe that such convergence warrants a central place in efforts to computationally emulate human cognition. That is, DCD is a general principle of cognition. It is also important to note that while we are drawing parallels across many literatures, these are functional parallels and are not necessarily structural ones. That is, we are not saying that the same neural pathways are involved in these phenomena. We are only saying that the different neural pathways that are responsible for the appearance of these various phenomena follow the same functional rules - the mechanisms are the same even if the physical parts are distinct. Furthermore, DCD is not a causal mechanism - it is an emergent property of the way the brain is constructed. DCD is the result of neurophysiology (cf. John, 2002, 2003). Finally, it is important to note that we are not proposing a generic learning mechanism such that one biological algorithm can account for all situation interpretation. Rather, we are pointing out that there are strikingly similar empirical results across a wide variety of disciplines that can be understood, in part, by similar cognitive processes. It is entirely possible, even assumed in some cases (i.e., primary language acquisition) that these more generic cognitive processes are complemented and constrained by various limits which may or may not be biological in nature (cf. Bates & Elman, 1996; Elman, in press).
As part of DARPA Information Processing Technology Office (IPTO) Software for Distributed Robotics (SDR) Program, Sandia National Laboratories has developed analysis and control software for coordinating tens to thousands of autonomous cooperative robotic agents (primarily unmanned ground vehicles) performing military operations such as reconnaissance, surveillance and target acquisition; countermine and explosive ordnance disposal; force protection and physical security; and logistics support. Due to the nature of these applications, the control techniques must be distributed, and they must not rely on high bandwidth communication between agents. At the same time, a single soldier must easily direct these large-scale systems. Finally, the control techniques must be provably convergent so as not to cause undo harm to civilians. In this project, provably convergent, moderate communication bandwidth, distributed control algorithms have been developed that can be regulated by a single soldier. We have simulated in great detail the control of low numbers of vehicles (up to 20) navigating throughout a building, and we have simulated in lesser detail the control of larger numbers of vehicles (up to 1000) trying to locate several targets in a large outdoor facility. Finally, we have experimentally validated the resulting control algorithms on smaller numbers of autonomous vehicles.
We present a model for optimizing the placement of sensors in municipal water networks to detect maliciously injected contaminants. An optimal sensor configuration minimizes the expected fraction of the population at risk. We formulate this problem as a mixed-integer program, which can be solved with generally available solvers. We find optimal sensor placements for three test networks with synthetic risk and population data. Our experiments illustrate that this formulation can be solved relatively quickly and that the predicted sensor configuration is relatively insensitive to uncertainties in the data used for prediction.
As Sandia looks toward petaflops computing and other advanced architectures, it is necessary to provide a programming environment that can exploit this additional computing power while supporting reasonable development time for applications. Thus, they evaluate the Partitioned Global Address Space (PGAS) programming model as implemented in Unified Parallel C (UPC) for its applicability. They report on their experiences in implementing sorting and minimum spanning tree algorithms on a test system, a Cray T3e, with UPC support. They describe several macros that could serve as language extensions and several building-block operations that could serve as a foundation for a PGAS programming library. They analyze the limitations of the UPC implementation available on the test system, and suggest improvements necessary before UPC can be used in a production environment.
The mathematical and physical foundations and domain of applicability of Sandia's GeoModel are presented along with descriptions of the source code and user instructions. The model is designed to be used in conventional finite element architectures, and (to date) it has been installed in five host codes without requiring customizing the model subroutines for any of these different installations. Although developed for application to geological materials, the GeoModel actually applies to a much broader class of materials, including rock-like engineered materials (such as concretes and ceramics) and even to metals when simplified parameters are used. Nonlinear elasticity is supported through an empirically fitted function that has been found to be well-suited to a wide variety of materials. Fundamentally, the GeoModel is a generalized plasticity model. As such, it includes a yield surface, but the term 'yield' is generalized to include any form of inelastic material response including microcrack growth and pore collapse. The geomodel supports deformation-induced anisotropy in a limited capacity through kinematic hardening (in which the initially isotropic yield surface is permitted to translate in deviatoric stress space to model Bauschinger effects). Aside from kinematic hardening, however, the governing equations are otherwise isotropic. The GeoModel is a genuine unification and generalization of simpler models. The GeoModel can employ up to 40 material input and control parameters in the rare case when all features are used. Simpler idealizations (such as linear elasticity, or Von Mises yield, or Mohr-Coulomb failure) can be replicated by simply using fewer parameters. For high-strain-rate applications, the GeoModel supports rate dependence through an overstress model.
One problem facing today's nuclear power industry is flow-accelerated corrosion and erosion in pipe elbows. The Korean Atomic Energy Research Institute (KAERI) is performing experiments in their Flow-Accelerated Corrosion (FAC) test loop to better characterize these phenomena, and develop advanced sensor technologies for the condition monitoring of critical elbows on a continuous basis. In parallel with these experiments, Sandia National Laboratories is performing Computational Fluid Dynamic (CFD) simulations of the flow in one elbow of the FAC test loop. The simulations are being performed using the FLUENT commercial software developed and marketed by Fluent, Inc. The model geometry and mesh were created using the GAMBIT software, also from Fluent, Inc. This report documents the results of the simulations that have been made to date; baseline results employing the RNG k-e turbulence model are presented. The predicted value for the diametrical pressure coefficient is in reasonably good agreement with published correlations. Plots of the velocities, pressure field, wall shear stress, and turbulent kinetic energy adjacent to the wall are shown within the elbow section. Somewhat to our surprise, these indicate that the maximum values of both wall shear stress and turbulent kinetic energy occur near the elbow entrance, on the inner radius of the bend. Additional simulations were performed for the same conditions, but with the RNG k-e model replaced by either the standard k-{var_epsilon}, or the realizable k-{var_epsilon} turbulence model. The predictions using the standard k-{var_epsilon} model are quite similar to those obtained in the baseline simulation. However, with the realizable k-{var_epsilon} model, more significant differences are evident. The maximums in both wall shear stress and turbulent kinetic energy now appear on the outer radius, near the elbow exit, and are {approx}11% and 14% greater, respectively, than those predicted in the baseline calculation; secondary maxima in both quantities still occur near the elbow entrance on the inner radius. Which set of results better reflects reality must await experimental corroboration. Additional calculations demonstrate that whether or not FLUENT's radial equilibrium pressure distribution option is used in the PRESSURE OUTLET boundary condition has no significant impact on the flowfield near the elbow. Simulations performed with and without the chemical sensor and associated support bracket that were present in the experiments demonstrate that the latter have a negligible influence on the flow in the vicinity of the elbow. The fact that the maxima in wall shear stress and turbulent kinetic energy occur on the inner radius is therefore not an artifact of having introduced the sensor into the flow.
The peridynamic theory of continuum mechanics allows damage, fracture, and long-range forces to be treated as natural components of the deformation of a material. In this paper, the peridynamic approach is applied to small thickness two- and one-dimensional structures. For membranes, a constitutive model is described appropriate for rubbery sheets that can form cracks. This model is used to perform numerical simulations of the stretching and dynamic tearing of membranes. A similar approach is applied to one-dimensional string like structures that undergrow stretching, bending, and failure. Long-range forces similar to van der Waals interactions at the nanoscale influence the equilibrium configurations of these structures, how they deform, and possibly self-assembly.
We are extending the existing features of Aspen, a powerful economic modeling tool, and introducing new features to simulate the role of confidence in economic activity. The new model is built from a collection of autonomous agents that represent households, firms, and other relevant entities like financial exchanges and governmental authorities. We simultaneously model several interrelated markets, including those for labor, products, stocks, and bonds. We also model economic tradeoffs, such as decisions of households and firms regarding spending, savings, and investment. In this paper, we review some of the basic principles and model components and describe our approach and development strategy for emulating consumer, investor, and business confidence. The model of confidence is explored within the context of economic disruptions, such as those resulting from disasters or terrorist events.
This document is a reference guide for the UNIX Library/Standalone version of the Latin Hypercube Sampling Software. This software has been developed to generate Latin hypercube multivariate samples. This version runs on Linux or UNIX platforms. This manual covers the use of the LHS code in a UNIX environment, run either as a standalone program or as a callable library. The underlying code in the UNIX Library/Standalone version of LHS is almost identical to the updated Windows version of LHS released in 1998 (SAND98-0210). However, some modifications were made to customize it for a UNIX environment and as a library that is called from the DAKOTA environment. This manual covers the use of the LHS code as a library and in the standalone mode under UNIX.
A study was undertaken to validate the 'capability' computing needs of DOE's Office of Science. More than seventy members of the community provided information about algorithmic scaling laws, so that the impact of having access to Petascale capability computers could be assessed. We have concluded that the Office of Science community has described credible needs for Petascale capability computing.
Two heuristic strategies intended to enhance the performance of the generalized global basis (GGB) method [H. Waisman, J. Fish, R.S. Tuminaro, J. Shadid, The Generalized Global Basis (GGB) method, International Journal for Numerical Methods in Engineering 61(8), 1243-1269] applied to nonlinear systems are presented. The standard GGB accelerates a multigrid scheme by an additional coarse grid correction that filters out slowly converging modes. This correction requires a potentially costly eigen calculation. This paper considers reusing previously computed eigenspace information. The GGB? scheme enriches the prolongation operator with new eigenvectors while the modified method (MGGB) selectively reuses the same prolongation. Both methods use the criteria of principal angles between subspaces spanned between the previous and current prolongation operators. Numerical examples clearly indicate significant time savings in particular for the MGGB scheme.
We consider linear systems arising from the use of the finite element method for solving scalar linear elliptic problems. Our main result is that these linear systems, which are symmetric and positive semidefinite, are well approximated by symmetric diagonally dominant matrices. Our framework for defining matrix approximation is support theory. Significant graph theoretic work has already been developed in the support framework for preconditioners in the diagonally dominant case, and in particular it is known that such systems can be solved with iterative methods in nearly linear time. Thus, our approximation result implies that these graph theoretic techniques can also solve a class of finite element problems in nearly linear time. We show that the support number bounds, which control the number of iterations in the preconditioned iterative solver, depend on mesh quality measures but not on the problem size or shape of the domain.
Analytic solutions are useful for code verification. Structural vibration codes approximate solutions to the eigenvalue problem for the linear elasticity equations (Navier's equations). Unfortunately the verification method of 'manufactured solutions' does not apply to vibration problems. Verification books (for example [2]) tabulate a few of the lowest modes, but are not useful for computations of large numbers of modes. A closed form solution is presented here for all the eigenvalues and eigenfunctions for a cuboid solid with isotropic material properties. The boundary conditions correspond physically to a greased wall.
It seems well understood that supercomputer simulation is an enabler for scientific discoveries, weapons, and other activities of value to society. It also seems widely believed that Moore's Law will make progressively more powerful supercomputers over time and thus enable more of these contributions. This paper seeks to add detail to these arguments, revealing them to be generally correct but not a smooth and effortless progression. This paper will review some key problems that can be solved with supercomputer simulation, showing that more powerful supercomputers will be useful up to a very high yet finite limit of around 1021 FLOPS (1 Zettaflops) . The review will also show the basic nature of these extreme problems. This paper will review work by others showing that the theoretical maximum supercomputer power is very high indeed, but will explain how a straightforward extrapolation of Moore's Law will lead to technological maturity in a few decades. The power of a supercomputer at the maturity of Moore's Law will be very high by today's standards at 1016-1019 FLOPS (100 Petaflops to 10 Exaflops), depending on architecture, but distinctly below the level required for the most ambitious applications. Having established that Moore's Law will not be that last word in supercomputing, this paper will explore the nearer term issue of what a supercomputer will look like at maturity of Moore's Law. Our approach will quantify the maximum performance as permitted by the laws of physics for extension of current technology and then find a design that approaches this limit closely. We study a 'multi-architecture' for supercomputers that combines a microprocessor with other 'advanced' concepts and find it can reach the limits as well. This approach should be quite viable in the future because the microprocessor would provide compatibility with existing codes and programming styles while the 'advanced' features would provide a boost to the limits of performance.
We compare inexact Newton and coordinate descent methods for optimizing the quality of a mesh by repositioning the vertices, where quality is measured by the harmonic mean of the mean-ratio metric. The effects of problem size, element size heterogeneity, and various vertex displacement schemes on the performance of these algorithms are assessed for a series of tetrahedral meshes.
There is currently a large research and development effort within the high-performance computing community on advanced parallel programming models. This research can potentially have an impact on parallel applications, system software, and computing architectures in the next several years. Given Sandia's expertise and unique perspective in these areas, particularly on very large-scale systems, there are many areas in which Sandia can contribute to this effort. This technical report provides a survey of past and present parallel programming model research projects and provides a detailed description of the Partitioned Global Address Space (PGAS) programming model. The PGAS model may offer several improvements over the traditional distributed memory message passing model, which is the dominant model currently being used at Sandia. This technical report discusses these potential benefits and outlines specific areas where Sandia's expertise could contribute to current research activities. In particular, we describe several projects in the areas of high-performance networking, operating systems and parallel runtime systems, compilers, application development, and performance evaluation.
We give processor-allocation algorithms for grid architectures, where the objective is to select processors from a set of available processors to minimize the average number of communication hops. The associated clustering problem is as follows: Given n points in R{sup d}, find a size-k subset with minimum average pairwise L{sub 1} distance.We present a natural approximation algorithm and show that it is a 7/4-approximation for 2D grids. In d dimensions, the approximation guarantee is 2 - 1/2d, which is tight. We also give a polynomial-time approximation scheme (PTAS) for constant dimension d and report on experimental results.
The Trilinos{trademark} Project is an effort to facilitate the design, development, integration and ongoing support of mathematical software libraries. AztecOO{trademark} is a package within Trilinos that enables the use of the Aztec solver library [19] with Epetra{trademark} [13] objects. AztecOO provides access to Aztec preconditioners and solvers by implementing the Aztec 'matrix-free' interface using Epetra. While Aztec is written in C and procedure-oriented, AztecOO is written in C++ and is object-oriented. In addition to providing access to Aztec capabilities, AztecOO also provides some signficant new functionality. In particular it provides an extensible status testing capability that allows expression of sophisticated stopping criteria as is needed in production use of iterative solvers. AztecOO also provides mechanisms for using Ifpack [2], ML [20] and AztecOO itself as preconditioners.
Sundance is a system of software components that allows construction of an entire parallel simulator and its derivatives using a high-level symbolic language. With this high-level problem description, it is possible to specify a weak formulation of a PDE and its discretization method in a small amount of user-level code; furthermore, because derivatives are easily available, a simulation in Sundance is immediately suitable for accelerated PDE-constrained optimization algorithms. This paper is a tutorial for setting up and solving linear and nonlinear PDEs in Sundance. With several simple examples, we show how to set up mesh objects, geometric regions for BC application, the weak form of the PDE, and boundary conditions. Each example then illustrates use of an appropriate solver and solution visualization.
This document is intended to contain a detailed description of the mathematical formulation of Xyce, a massively parallel SPICE-style circuit simulator developed at Sandia National Laboratories. The target audience of this document are people in the role of 'service provider'. An example of such a person would be a linear solver expert who is spending a small fraction of his time developing solver algorithms for Xyce. Such a person probably is not an expert in circuit simulation, and would benefit from an description of the equations solved by Xyce. In this document, modified nodal analysis (MNA) is described in detail, with a number of examples. Issues that are unique to circuit simulation, such as voltage limiting, are also described in detail.
This SAND report is the final report on Sandia's Grand Challenge LDRD Project 27328, 'A Revolution in Lighting -- Building the Science and Technology Base for Ultra-Efficient Solid-state Lighting.' This project, which for brevity we refer to as the SSL GCLDRD, is considered one of Sandia's most successful GCLDRDs. As a result, this report reviews not only technical highlights, but also the genesis of the idea for Solid-state Lighting (SSL), the initiation of the SSL GCLDRD, and the goals, scope, success metrics, and evolution of the SSL GCLDRD over the course of its life. One way in which the SSL GCLDRD was different from other GCLDRDs was that it coincided with a larger effort by the SSL community - primarily industrial companies investing in SSL, but also universities, trade organizations, and other Department of Energy (DOE) national laboratories - to support a national initiative in SSL R&D. Sandia was a major player in publicizing the tremendous energy savings potential of SSL, and in helping to develop, unify and support community consensus for such an initiative. Hence, our activities in this area, discussed in Chapter 6, were substantial: white papers; SSL technology workshops and roadmaps; support for the Optoelectronics Industry Development Association (OIDA), DOE and Senator Bingaman's office; extensive public relations and media activities; and a worldwide SSL community website. Many science and technology advances and breakthroughs were also enabled under this GCLDRD, resulting in: 55 publications; 124 presentations; 10 book chapters and reports; 5 U.S. patent applications including 1 already issued; and 14 patent disclosures not yet applied for. Twenty-six invited talks were given, at prestigious venues such as the American Physical Society Meeting, the Materials Research Society Meeting, the AVS International Symposium, and the Electrochemical Society Meeting. This report contains a summary of these science and technology advances and breakthroughs, with Chapters 1-5 devoted to the five technical task areas: 1 Fundamental Materials Physics; 2 111-Nitride Growth Chemistry and Substrate Physics; 3 111-Nitride MOCVD Reactor Design and In-Situ Monitoring; 4 Advanced Light-Emitting Devices; and 5 Phosphors and Encapsulants. Chapter 7 (Appendix A) contains a listing of publications, presentations, and patents. Finally, the SSL GCLDRD resulted in numerous actual and pending follow-on programs for Sandia, including multiple grants from DOE and the Defense Advanced Research Projects Agency (DARPA), and Cooperative Research and Development Agreements (CRADAs) with SSL companies. Many of these follow-on programs arose out of contacts developed through our External Advisory Committee (EAC). In h s and other ways, the EAC played a very important role. Chapter 8 (Appendix B) contains the full (unedited) text of the EAC reviews that were held periodically during the course of the project.