Abstract: A method and apparatus for estimating and/or controlling mercury emissions in a steam generating unit. A model of the steam generating unit is used to predict mercury emissions. In one embodiment of the invention, the model is a neural network (NN) model. An optimizer may be used in connection with the model to determine optimal setpoint values for manipulated variables associated with operation of the steam generating unit.

Abstract: An information processing apparatus for configuring algorithms is disclosed. The information processing apparatus includes an algorithm configuring section that configures an algorithm for performing discrimination on an input signal by using a genetic search technique. The algorithm includes feature extraction expressions and an information estimation expression represented by a combination of the feature extraction expressions. The information processing apparatus also includes a tradeoff analyzing section that determines pareto optimal solutions by optimizing the algorithm with respect to evaluation indices by performing tradeoff analysis on the basis of the algorithm. In addition, the information processing apparatus includes a storage for storing the algorithm.

Abstract: In various energy systems which handle a plurality of types of energy, a method and device for operation scheduling for energy storage equipment are provided, which determines optimal operation of an energy storage equipment and improves the efficiency of the energy system overall. An energy storage equipment operation scheduling part 13 creates an energy storage equipment operation schedule, an energy generation equipment modified operation schedule, and an energy storage equipment modified operation schedule, based on three information items, which are the energy demand forecast information stored in a storage part D1, the energy generation equipment operation schedule stored in a storage part D2, and equipment connection information, and sends the created schedules to the equipment controller 20.

Abstract: Systems and method are disclosed for query optimization in a scale-out system with a single query processing machine and a distributed storage engine to store data by receiving a query rewritten for an internal schema; optimizing a query execution plan for the query; and executing the plan and returning result to an application.

Abstract: The robust controller for nonlinear MIMO systems uses a radial basis function (RBF) neural network to generate optimal control signals abiding by constraints, if any, on the control signal or on the system output. The weights of the neural network are trained in the negative direction of the gradient of output squared error. Nonlinearities in the system, as well as variations in system parameters, are handled by the robust controller. Simulation results are included in the end to assess the performance of the proposed controller.

Abstract: According to an embodiment of the present invention, a system optimizes an information processing environment, and comprises at least one processor. The system collects information pertaining to operational behavior of the information processing environment and including a plurality of parameters. A neural network structure is established to associate the parameters to a desired operational performance characteristic for the information processing environment. The neural network structure is trained with the collected information from the information processing environment to produce a model for the information processing environment. The model is optimized to determine values for the parameters and the information processing environment is adjusted based on the determined parameter values to attain the desired operational performance of the information processing environment.

Abstract: Upon detecting that a point that satisfies a predetermined condition and whose distance from a reference point is shorter than a distance from the reference point to a first point obtained by searching a parameter space based on values of a first search indicator under a first constraint exists in the parameter space, a second point is calculated under the first constraint in the parameter space by a search method other than the searching using the first search indicator. Then, generating a second search indicator represented by a first linear combination of at least certain of first search indicators so as to obtain the second point or an adjacent point of the second point, when searching by using the second search indicator or generating a second search indicator so that search is carried out in a direction of the second point when using the second search indicator is carried out.

Abstract: Approaches for accurate neural network training for library-based critical dimension (CD) metrology are described. Approaches for fast neural network training for library-based CD metrology are also described. In an example, a method includes optimizing a threshold for a principal component analysis (PCA) of a spectrum data set to provide a principal component (PC) value, estimating a training target for one or more neural networks, training the one or more neural networks based both on the training target and on the PC value provided from optimizing the threshold for the PCA, and providing a spectral library based on the one or more trained neural networks.

Abstract: A novel eco-system is provided which first supplies a standardized template of one or more virtual machine images for software module providers/vendors. A plurality of modules executing on the virtual machine images is selected by a user to comprise a plurality of configurations. A suitable configuration may be determined according to a metric and the determined suitable configuration of software modules is subsequently used to build an end-to-end solution.

Abstract: A system and method for deciding the satisfiability of a non-linear real decision problem is disclosed. Linear and non-linear constraints associated with the problem are separated. The feasibility of the linear constraints is determined using a linear solver. The feasibility of the non-linear constraints is determined using a non-linear solver which employs interval constraint propagation. The interval solutions obtained from the non-linear solver are validated using the linear solver. If the solutions cannot be validated, linear constraints are learned to refine a search space associated with the problem. The learned constraints and the non-linear constraints are iteratively solved using the non-linear solver until either a feasible solution is obtained or no solution is possible.

Abstract: Embodiments of the present invention describe using a data structure to represent variable domains in solving a constraint problem. The data structure includes nodes that are configured to represent ranges of values in order to save memory space and processor power. Ranges of values and single values may be both added to and removed from the data structure such that the data structure does not include duplicate values. Operations may include detecting multiple nodes with adjacent or overlapping ranges that may be combined into a single node, and adding or removing all the values in the structure that are beyond a threshold value. In one embodiment the data structure may be a balanced binary tree. Constraint solvers may either add values to the data structure as the values are eliminated from the solution, or they may remove those values.

Abstract: A method for predicting and optimizing magnetic core width of a write head using neural networks to analyze manufacturing parameters, and determining new manufacturing parameters that will provide more optimal magnetic core width results. The manufacturing parameters can include: write pole flare point; wrap around shield dimension; and side gap dimension.

Abstract: An approach for solving a global optimization problem is described. Specifically, one embodiment of the disclosure sets forth a method, which includes the steps of receiving a quantitative initial solution, generating a quantitative feasible solution, mapping the quantitative feasible solution to a qualitative feasible solution, determining whether to accept the qualitative feasible solution based on a first predetermined rule, wherein the qualitative feasible solution that is accepted is reverse mapped to the quantitative feasible solution, and transmitting a result of the determining step.

Abstract: A system for solving a dynamic constraint satisfaction problem comprises a constraint network of variables and constraints. The system creates a first sub-problem model that includes a first model type, one or more first variables and zero or more first constraints. The system propagates the first constraints through the constraint network and determines if a first conflict is detected from propagating the first constraints. If the first conflict is detected, the system restores the constraint network variables to a first previous state before the first constraints were propagated. The system creates a first sub-problem set that includes a second model type and one or more sub-problem models. The system connects the first sub-problem model to the first sub-problem set via a second constraint and propagates the second constraint through the constraint network.

Abstract: A computer-implemented land planning system is designed to generate at least one conceptual fit solution to a user-defined land development problem. The system electronically creates at least one candidate solution to the land development problem. The candidate solution incorporates a number of engineering measurements applicable in development of an undeveloped land site. A fitness function quantitatively evaluates the candidate solution based on its fitness. A heuristic problem-solving strategy manipulates the engineering measurements of the candidate solution to achieve a more quantitatively fit solution to the land development problem. A computer output device outputs to a user documentation illustrating the fit solution to the land development problem.

Abstract: Managing a portfolio of experts is described where the experts may be for example, automated experts or human experts. In an embodiment a selection engine selects an expert from a portfolio of experts and assigns the expert to a specified task. For example, the selection engine has a Bayesian machine learning system which is iteratively updated each time an experts performance on a task is observed. For example, sparsely active binary task and expert feature vectors are input to the selection engine which maps those feature vectors to a multi-dimensional trait space using a mapping learnt by the machine learning system. In examples, an inner product of the mapped vectors gives an estimate of a probability distribution over expert performance. In an embodiment the experts are automated problem solvers and the task is a hard combinatorial problem such as a constraint satisfaction problem or combinatorial auction.

Abstract: A computer readable storage medium includes executable instructions to derive from a database schema an irreducible ambiguous group comprising a sub-schema with a set of vertices wherein any two vertices are part of a loop. Contexts are defined on the sub-schema. For each context, joins in the sub-schema are designated as mandatory joins, excluded joins and neutral joins. A selection of a context from multiple contexts invoked by a path characterizing a query is processed. The query is resolved using the context.

Abstract: Embodiments of the disclosed systems and methods establish quantitative relationships between system features and system objectives. In some embodiments, the features have a plurality of feature values related to the objective and the methods comprise analyzing a mathematical functional relationship between the plurality of feature values and the objective to create a plurality of objective values reflecting the ability of the feature values to satisfy the objective, selecting a feature value and analyzing the relationship to create an objective value; and generating an objective measure reflecting the objective value. In some embodiments, the mathematical function comprises a polynomial interpolation. In some embodiments, the features are a fidelity dimension and the feature values are values of fidelity in a processor based aircraft simulator.

Abstract: The robust controller for nonlinear MIMO systems uses a radial basis function (RBF) neural network to generate optimal control signals abiding by constraints, if any, on the control signal or on the system output. The weights of the neural network are trained in the negative direction of the gradient of output squared error. Nonlinearities in the system, as well as variations in system parameters, are handled by the robust controller. Simulation results are included in the end to assess the performance of the proposed controller.

Abstract: A method for predicting results for input data based on a model that is generated based on clusters of related characters, clusters of related segments, and training data. The method comprises receiving a data set that includes a plurality of words in a particular language. In the particular language, words are formed by characters. Clusters of related characters are formed from the data set. A model is generated based at least on the clusters of related characters and training data. The model may also be based on the clusters of related segments. The training data includes a plurality of entries, wherein each entry includes a character and a designated result for said character. A set of input data that includes characters that have not been associated with designated results is received. The model is applied to the input data to determine predicted results for characters within the input data.

Abstract: Systems and methods for predicting energy usage of an asset are provided. Among several implementations of methods implemented by a computer, one embodiment of a computer-implemented method includes selecting one of a plurality of base temperatures that allows a linear equation to estimate energy consumption by an asset as a function of an average daily demand on the asset to attain a desired temperature. The computer-implemented method also includes inserting the selected base temperature in a non-linear equation for modeling the asset's energy consumption.

Abstract: Approaches for performing simulation optimization for solving a constrained optimization problem are generally disclosed. One embodiment according to the present disclosure is to formulate a Lagrange equation having incorporated a Lagrange parameter, a first long run average function for an objective associated with the constrained optimization problem, and a second long run average function for a constraint associated with the constrained optimization problem. Then, to identify a parameter value that may lead to an extreme value for the Lagrange equation, in an iterative manner, averages of the first long run average function and the second long run average function are calculated, a gradient of the Lagrange equation is estimated, and the Lagrange parameter is updated.

Abstract: A classification system is described for resilient classification of data. In various embodiments, the classification system divides a full set of the training data into a positive subset and a negative subset. The positive subset contains all training data with a positive classification value and the negative subset contains all training data with a negative classification value. The classification system constructs several subsets of the training data wherein each subset contains data randomly selected from both the positive subset and the negative subset. The classification system then creates at least two classifiers based on each of the randomly created subsets of the training data.

Abstract: Systems and associated methods provide a cluster-level semi-supervision model for inter-active clustering. Embodiments accept user provided semi-supervision for updating cluster descriptions and assignment of data items to clusters. Assignment feedback re-assigns data items among existing clusters, while cluster description feedback helps to position existing cluster centers more meaningfully. The feedback can continue until the user is satisfied with the clustering achieved or one or more predetermined stopping criteria have been reached.

Abstract: Certain aspects of the present disclosure support a local competitive learning rule applied in a computational network that leads to sparse connectivity among processing units of the network. The present disclosure provides a modification to the Oja learning rule, modifying the constraint on the sum of squared weights in the Oja rule. This constraining can be intrinsic and local as opposed to the commonly used multiplicative and subtractive normalizations, which are explicit and require the knowledge of all input weights of a processing unit to update each one of them individually. The presented rule provides convergence to a weight vector that is sparser (i.e., has more zero elements) than the weight vector learned by the original Oja rule. Such sparse connectivity can lead to a higher selectivity of processing units to specific features, and it may require less memory to store the network configuration and less energy to operate it.

Abstract: Real-time condition-based analysis is performed on a machine for providing diagnostic and prognostic outputs indicative of machine status includes a signal processor for receiving signals from sensors adapted for measuring machine performance parameters. The signal processor conditions and shapes at least some of the received signals into an input form for a neural network. A fuzzy adaptive resonance theory neural network receives at least some of the conditioned and shaped signals, and detects and classifies a state of the machine based upon the received conditioned and shaped signals, and upon a predetermined ontology of machine states, diagnostics, and prognostics. The neural network can also determine from the machine state a health status thereof, which can comprise an anomaly, and output a signal representative of the determined health status. A Bayesian intelligence network receives the machine state from the neural network and determines a fault probability at a future time.

Abstract: A system for determining an optimal forecasting hierarchy includes setting a plurality of hierarchies by setting each hierarchy in terms of a hierarchy structure. The system also includes generating top level forecasts for a training period and a testing period, each top level forecast being based upon an aggregation of the base forecast levels. The system further includes calculating a forecast error. An optimization routine includes determining an optimal hierarchy and an optimal base level, both being based upon a smallest associated forecast error. The optimization routine also includes calculating a deficiency ratio for each level within each hierarchy by comparing forecast errors associated with the optimal base level and the optimal hierarchy to forecast errors associated with each of the other base levels and hierarchies. The optimal forecasting hierarchy is determined by comparing the deficiency ratio to a significance level.

Abstract: A method for predicting and optimizing magnetic core width of a write head using neural networks to analyze manufacturing parameters, and determining new manufacturing parameters that will provide more optimal magnetic core width results. The manufacturing parameters can include: write pole flare point; wrap around shield dimension; and side gap dimension.

Abstract: A novel method of automated, real-time website optimization at least includes: a) receiving website optimization data including an optimization goal, and website source code; b) receiving website optimization criteria indicative of the completion of a website optimization experiment; c) executing an optimization algorithm used to select an optimized website version; d) comparing the output of the optimization algorithm with the website optimization goal to determine whether the website version under consideration is optimized; e) providing feedback from the output of the executed optimization algorithm to an input of the optimization algorithm; f) based upon the feedback, determining the next iterative step of the optimization algorithm; g) performing new iterative steps of the optimization algorithm; h) converging to an optimized website state; and i) modifying the website source code to implement the optimum version of the website.

Abstract: A computer-implemented land planning system is designed to generate at least one conceptual fit solution to a user-defined land development problem. The system electronically creates at least one candidate solution to the land development problem. The candidate solution incorporates a number of engineering measurements applicable in development of an undeveloped land site. A fitness function quantitatively evaluates the candidate solution based on its fitness. A heuristic problem-solving strategy manipulates the engineering measurements of the candidate solution to achieve a more quantitatively fit solution to the land development problem. A computer output device outputs to a user documentation illustrating the fit solution to the land development problem.

Abstract: A system and method provides a solution to the problem of applying end-to-end requirements of connectivity, security, reliability and performance to configure a network and ultimately assign network components to the network. All requirements are modeled as constraints and a constraint solver does the resolution. Not every constraint to be solved is solved by the model-finder. Instead, we “factor away” subsets of a constraint that can be efficiently solved via a special-purpose constraint solver, such as an SQL/Prolog engine, linear programming system, or even an algorithm, leaving behind a constraint that truly requires the power of model-finding, and that is often efficiently solvable by existing model-finders. Such constraints are compiled into quantifier-free constraints that are Boolean combinations of constraints of two forms x=y and x=c where x, y are variables and c is a constant. Such constraints can be efficiently solved by modern SAT-based model-finders.

Abstract: A method of scheduling availability for a computing infrastructure in a shared computing environment is disclosed. The method comprises assigning a new schedule of service to a software application in the computing environment, propagating the new schedule of service assigned to the software application to a plurality of computing components that support the software application, where the computing infrastructure is comprised of the plurality of computing components having a hierarchical relationship with each other, and determining a schedule of service for a given computing component in the computing infrastructure by aggregating schedules of service propagated to the given computing component.

Abstract: The present invention provides a method for verification of linear hybrid automaton by generation an initial abstract model based on an original Linear-Time Temporal Logic (LTL) specification, validating a counterexample using an approach of linear constraints, identifying a fragment in the counterexample by iteratively applying an approach of linear constraints satisfaction in a limited number of times, and refining the original LTL specification based on the fragment derived.

Abstract: The present invention is a method of solving the decision, for example, testing if, given a finite number of transformations which can be applied to a finite number of elements, the corresponding n-generated discrete object has a hamiltonian cycle and/or path, searching, for example, obtaining the explicit construction of one several or all the hamiltonian cycles and or paths of the given input, counting, for example, obtaining an upper bound of the number of Hamiltonian cycles and/or paths of the given input and optimization, for example, selecting one of several hamiltonian cycles and/or paths solutions according to an specified criterion, versions of the hamiltonian traversal (cycle and/or path) problem in class of combinatorial discrete objects.

Abstract: Approaches for accurate neural network training for library-based critical dimension (CD) metrology are described. Approaches for fast neural network training for library-based CD metrology are also described.

Abstract: A computer-implemented land planning system is designed to generate at least one conceptual fit solution to a user-defined land development problem. The system electronically creates at least one candidate solution to the land development problem. The candidate solution incorporates a number of engineering measurements applicable in development of an undeveloped land site. A fitness function quantitatively evaluates the candidate solution based on its fitness. A heuristic problem-solving strategy manipulates the engineering measurements of the candidate solution to achieve a more quantitatively fit solution to the land development problem. A computer output device outputs to a user documentation illustrating the fit solution to the land development problem.

Abstract: A “variable group selection” system and method in which constructs are based upon a training data set, a regression modeling module that takes into account information on groups of related predictor variables given as input and outputs a regression model with selected variable groups. Optionally, the method can be employed as a component in methods of temporal causal modeling, which are applied on a time series training data set, and output a model of causal relationships between the multiple times series in the data.

Abstract: A method for generating precedence-preserving crossover and mutations operations for genetic algorithms is provided. The method is based on the determination of activities' Forward Free Float (FFF) and Backward Free Float (BFF) values, utilizing these float values in randomly selected forward and backward paths, respectively. The method may be applied to the finance-based scheduling domain using large scale projects, with the chromosomes of the genetic algorithm encoding activities' start times in a resource-constrained scheduling problem.

Abstract: Methods, systems, and apparatus, including computer programs encoded on one or more computer storage devices, for training and retraining predictive models. A series of training data sets for predictive modeling can be received, e.g., over a network from a client computing system. The training data included in the training data sets is different from initial training data that was used with multiple training functions to train multiple trained predictive models stored in a predictive model repository. The series of training data sets are used with multiple trained updateable predictive models obtained from the predictive model repository and multiple training functions to generate multiple retrained predictive models. An effectiveness score is generated for each of the retrained predictive models. A first trained predictive model is selected from among the trained predictive models included in the predictive model repository and the retrained predictive models based on their respective effectiveness scores.

Abstract: An analysis tool for causing a computer to use information gain of attributes and a classification algorithm to classify new records in a set of data by taking into account the predictive value of the attributes and the effect of the new record.

Abstract: A solver for a constraint satisfaction problem includes a plurality of variables and a plurality of constraints. A floating point variable has a domain and is assigned a value by first determining if a predetermined value can be assigned to the floating point variable if the predetermined value is in the domain. If not, the solver determines if a bound of the domain can be assigned to the floating point variable. If the predetermined value can not be assigned to the floating point variable and the bound of the domain can not be assigned to the floating point variable, the solver assigns a value to the floating point variable using domain splitting.

Abstract: Systems and methods for managing floating point variables are described in the present disclosure. According to one example, an embodiment of a method is described. The method comprises providing a floating point variable having a domain that includes a flag representing whether a specific value is included in or excluded from the domain of the floating point variable. The method also includes analyzing a constraint on the floating point variable to determine if the constraint excludes the specific value from the domain of the floating point variable. A value of the flag is manipulated to indicate that the specific value is excluded from the domain of the floating point variable if it is determined that the constraint excludes the specific value. In some cases, the specific value can be the value zero, for example.

Abstract: Nonlinear optimization is applied to resource allocation, as for example, buffer pool optimization in computer database software where only the marginal utility is known. The method for allocating resources comprises the steps of starting from an initial allocation, calculating the marginal utility of the allocation, calculating the constraint functions of the allocation, and applying this information to obtain a next allocation and repeating these steps until a stopping criteria is satisfied, in which case a locally optimal allocation is returned.

Abstract: Components of a distributed computing system are monitored, the components including hardware components and software components that operate on the hardware components. At least one of the software components is a service that includes a service level agreement. Performance characteristics of the components are determined based on the monitoring. The performance characteristics of the service are compared to the service level agreement to determine whether the service level agreement has been violated. At least one of the service or an additional service collocated with the service is migrated based on the performance characteristics of the components if the service level agreement has been violated.

Abstract: Multi-Objective Optimization (MOO) is integrated with a Constraint Management System (CMS) so as to rapidly and flexibly search large design spaces and focus on “interesting” designs as determined by user-specified criteria. A method embeds a trade space and its trade-off envelope within the same CMS network used to calculate the values of the points in the trade space itself. The method supports automatic variation of the resulting trade-off analyses in response to variations in upstream parameters, assumptions, and/or requirements. When feasible, the CMS can back-solve for values of upstream parameters so that selected attributes of the trade-off envelope achieve user-specified values. The coupled use of a CMS with multi-objective optimization avoids having to generate a large set of permutations of inputs and factors and then restricting the analysis to those having desired outputs or of having to manually “reformulate” the equation set so that it solves in the desired direction.

Abstract: A method, and corresponding computer program product and system, defines and uses marker points within a modeled manufacturing process routing that includes multiple sequenced operations. The method includes receiving user input that defines one or more marker points within the modeled manufacturing process routing and between sequential ones of the operations. The marker points define a user-defined point within a manufacturing process and include one of multiple defined types that each define a different use to be made by the marker point. The method also includes detecting if any marker points of a specified one of the defined types have been defined in the manufacturing process routing. If a marker point having the specified one of the defined types is detected, a predefined computing function is executed that uses the detected marker point.

Abstract: A system, method, and computer-readable medium for optimizing execution of a join operation in a parallel processing system are provided. A plurality of processing nodes that have at least one row of one or more tables involved in a join operation are identified. For each of the processing nodes, respective counts of rows that would be redistributed to each of the processing nodes based on join attributes of the rows are determined. A redistribution matrix is calculated from the counts of rows of each of the processing nodes. An optimized redistribution matrix is generated from the redistribution matrix, wherein the optimized redistribution matrix provides a minimization of rows to be redistributed among the nodes to execute the join operation.

Abstract: In an embodiment, a constraint is created for a database table. The constraint specifies a condition for a first column in the database table and an action. The action specifies whether data that violates the condition is allowed to be stored in the first column. A value and a specification of a second column in the database table are received from a data source. If the second column is identical to the first column, the value violates the condition, and the action specifies that data that violates the condition is allowed to be stored, the value is stored in a row in the database, the row is marked as hidden, and an identification of the constraint that was violated is stored in the row. A query does not return the row that is marked as hidden.

Abstract: Problem to be solved by the invention: The present invention aims at providing a motion discrimination system and a method for motion discrimination capable of real-time and accurate detecting motions of an object. Means to solve the problem: A motion discrimination system according to the present invention comprises: a detecting means for outputting different analog values corresponding to motion of an object in each of a plurality of detecting areas; a data processing means for processing the analog values output corresponding to the motion in each of the plurality of detecting areas to quantized time-series data; and a motion discrimination means for outputting a motion discrimination result using a motion discrimination model in which the time-series data are inputs and types of motions are outputs. The motion discrimination model is a neural network in which the time-series data are inputs to input nodes and the types of motions are outputs from output nodes. Selected drawing: FIG.

Abstract: A system and method for optimizing system performance includes applying sampling based optimization to identify optimal configurations of a computing system by selecting a number of configuration samples and evaluating system performance based on the samples. Based on feedback of evaluated samples, a location of an optimal configuration is inferred. Additional samples are generated towards the location of the inferred optimal configuration to further optimize a system configuration.