International Business Machines Corporation

Junya Shimizu

Kazuo Sekiya

Michihiro Yokoyama

Hiroaki Nakamura

Hiroshi Ishikawa

Tadanobu Inoue

Takashi Sakairi

Noboru Kamijo

Masayuki Murata

Fusashi Nakamura

Kazutoshi Sugimoto

Yuriko Sawatani

Ibuki Hara

Kosato Makita

Methods, computer program products, and systems are presented. The methods include, for instance: obtaining service life data of a part including a maintenance operation count of the part. The life expectancy of the part is formulated as a function of a life span of the part and the maintenance operation count at a point in time based on the obtained service life data and data interpolated therefrom. A life expectancy model is built based on the function and a plurality of life expectancies are predicted by applying simulated inputs to the life expectancy model. The life expectancies are produced after verification to indicate when to replace the part.

Predicting life expectancy of machine part

A method for interactive test-schedule adjustment is provided. The steps include: displaying a table including the importance and period of each test item; displaying a first optimum schedule obtained from an optimal solution based on values entered by a user into the table, the schedule including a use schedule of each test facilities used for each of the multiple test items; displaying a second optimum schedule obtained from an optimal solution recalculated by mathematical programming after a use schedule of a test item the importance of which is changed on the first optimum schedule is changed; and displaying a third optimum schedule based on neighborhood solutions obtained by a search method using a history of optimal solutions if no agreement between the users is obtained on the second schedule.

Interactive test-schedule adjustment method

Methods, computer program products, and systems are presented. The methods include, for instance: modeling for a subject process by machine learning with adaptive inputs. In one embodiment, the modeling may include: generating a model by use of machine learning with training data from measurements of successive components of a process to be modeled in order to predict measurements of a succeeding component within a statistically meaningful prediction range; adjusting the generated model by use of machine learning with less-deviation inducing measurements from a preceding component in case the measurement of the succeeding component is out of the prediction range; and presenting the adjusted model as a prediction model for the process.

Modeling a subject process by machine learning with adaptive inputs

A system, program and related computer implemented method for selecting a controller and constructing it to automate one or more processes is provided. The method comprises the steps of: constructing a controller from a plant described by an LPV model to which a scheduling parameter is given by: a) calculating one or more &#x3bd; gaps and a mean value of each candidate of a plant parameter vector of said LPV model based on a requirement of preselected constraints; b) choosing a candidate via information gathered by proof based on the magnitude of differences between the &#x3bd; gaps; c) converting the LPV model into LMI and finding controllers, according to an Hcriterion, at all combination end points of a maximum value and a minimum value of the plant parameter vector thus chosen; and d) choosing and constructing said chosen controller for said plant.

Method, system and program for constructing a controller

A system is configured so that signals passed through an all-pass filter using warp parameter &#x3bb; are whitening-filtered, and so that the frequency axis is restored by an all-pass filter using warp parameter &#x3bb;. This optimizes the whitening filter by determining the optimum &#x3bb;. First, the AR order p is automatically estimated using &#x3bb;=0. A spectral distance d&#x3bb; is computed using a discrete Fourier transform spectrum value passed through an all-pass filter using warp parameter &#x3bb; and a discrete AR spectrum value passed through an all-pass filter using warp parameter &#x3bb;. The &#x3bb; which minimizes the spectral distance d&#x3bb; is set as the warp parameter, and a whitening filter is configured in which the warp parameter has been optimized.

Whitening filter configuration method, program, and system

Method, article of manufacture and system using minimum data to determine whether a sliding-mode control should be applied in a plant. First measure the plant in an open-loop control fashion, and using the measured data, describe a state equation of the plant by system identification and order determination methods. Then design a switching hyperplane for sliding-mode control. Next, calculate higher order statistics on the difference between an output of a linear model on the hyperplane and an output of the sliding-mode control model in the measured data; When any of the higher order moments is larger than a predetermined threshold, configure a controller as a sum of the linear control input term and the nonlinear control input term. If both higher-order moments are smaller than the threshold, then configure the controller using only a linear control input term.

Method, article of manufacture, and system for configuring controller in sliding-mode control scheme

A method and system for processing Kalman Filter. The system includes: an Unscented Kalman Filter; and a processor device configured to: non-uniform a phase duration of a signal outputted from a plant; inputting the signal to the Unscented Kalman Filter; and restore non-uniformed phase duration of an estimated value calculated in the Unscented Kalman Filter to the phase duration.

Processing Kalman filter

A processing system, method, and program for a modeling system. A math equation including a differential equation is prepared. A CAS solves the differential equation normally and by using an interval coefficient in accordance with an interval analysis technique. A coefficient vector of the two solutions are compared to determine if a distance between the vectors is equal to or less than a preset value, and if so, the CAS solves the math equation including a differential equation with the target coefficient set to zero. The coefficient vectors of the solution thus obtained and the normally solved solution are compared to determine if the distance between the vectors is equal to or less than a preset value, and if so, the differential equation is simplified by omitting a term of the target coefficient from the differential equation.

Processing system, method, and program for modeling system

A method and system for configuring a PID controller. The method includes the steps of: providing initial values of a coefficient candidate group for a PID controller, where the coefficient candidate group includes coefficient candidates; providing an evaluation functional, where an unfalsified evaluation function is calculated from the evaluation functional; measuring an input and an output of a plant; selecting the coefficient candidates from the coefficient candidate group; calculating the evaluation functional using measured input and measured output for each of the coefficient candidates of the PID controller; removing a selected coefficient candidate if the evaluation functional is falsified as a result of a calculation of the evaluation functional; selecting new coefficient candidates as an update expression from remaining coefficient groups that have not been falsified; and repeating the steps of calculating and removing for each of the new coefficient candidates of an updated coefficient candidate group.

Configuring a PID controller

[Object] To generate a signal pattern generation apparatus capable of performing suitable processing even when an abrupt increase in the number of accesses occurs while the number of accesses is large, having an adequate long-tail characteristic, and capable of providing adequate fractality with respect to a plurality of ranges with respect to time and providing a suitable short-range distribution characteristic. [Solution] A signal pattern generation apparatus of the present invention is constituted by a root signal generation sectionwhich generates a root signal of a long-tail distribution having positivity, a statistical self-similarity signal generation section SSSSG which generates a statistical self-similarity signal based on wavelet analysis from the root signal, a general-filter sectionwhich decomposes the statistical self-similarity signal into at least two intermediate patterns, a coefficient acquisition sectionwhich obtains coefficients for the two intermediate patterns, and a resynthesis sectionwhich generates a resynthesized pattern from the two intermediate patterns and the obtained coefficients.

Signal pattern generation apparatus, signal pattern generation method, program for making computer system execute the signal pattern generation method, computer-readable storage medium on which the program is stored, network endurance testing system, and network endurance testing method

Systems and methods for distribution-transition estimation of key performance indicator (KPI). Exemplary embodiments include a method including establishing a state-space model in a computer memory, the model for handling measurement errors and modelling errors as noise, wherein time-varying indices have historical data and a correlation with KPI, adaptively estimating parameters related to the time varying indices, and obtaining residuals from the adaptive estimate of the parameters, estimating a probability distribution in order to generate random numbers from the probability distribution, generating random numbers related to the estimated probability distribution, thereby enabling prediction of future indicator distributions through a Sequential Monte Carlo method, providing models to input KPI intensity levels for a scenario level and to predict future values for a plurality of scenarios as point estimates and distributing and synthesizing the point estimation results for the plurality of scenarios based on KPI estimated probability distributions.

Systems and methods for distribution-transition estimation of key performance indicator

A method and system for determining a variable compensation amount according to the degree of achievement by avoiding stop at the stage where value brought about to a client by a solution is qualitatively evaluated, avoiding compensation calculation from being performed by accumulation of cost such as labor cost, and quantitatively estimating the effect from the viewpoint of a set KPI in consideration of risk to achievement of the KPI. There is provided a system for supporting advance evaluation of a service for improving business, comprising: input means for accepting selection of an index indicating the degree of improvement of the business and accepting setting of multiple future scenarios which influences success/failure of the improvement of the business; storage means for storing the index and the multiple scenarios; and calculation means for calculating a primary estimate value of the index in the case of occurrence of each scenario and calculating a secondary estimate value based on the primary estimate value and a primary risk corresponding to the primary estimate value.

Computer program for supporting evaluation of a service

It is an object of the present invention to output a clear, expanded image wherein step-shapes or chain-shapes of oblique lines are reduced, distortion is eliminated and no constituent lines of fonts or graphics are missing. An image transform method, for transforming original input image data into image data expanded by a ratio represented by a rational number or an integer, comprises: a vertical and horizontal interpolation unit, for reducing correlation in the vertical and horizontal directions of interpolated image data linearly expanded from an original image data input unit, and for generating first expanded image data; an oblique interpolation unit, for performing linear interpolation, based on correlation with a target pixel constituting the original image data and neighboring pixels arranged in oblique directions, using the neighboring pixels to generate second expanded image data; and a final result generator for employing the first expanded image data and the second expanded image data to generate a final image.

Image transform method for obtaining expanded image data, image processing apparatus and image display device therefore

Apparatus and methods for computer file management including components and steps for distributing to servers a plurality of mixed images that are generated for each of N, where N&#x2267;3, sets of distributed information generated by distributing secret information using a threshold secret sharing scheme, the plurality of mixed images being generated by mixing at different mixing ratios an image based on the distributed information with (Ka&#x2212;1) images, where Ka satisfies 1&#x3c;Ka&#x3c;N; transmitting Kb passwords, where Kb satisfies 1&#x3c;Kb&#x3c;N, entered by a user, and obtaining from the plurality of servers at least Ka mixed images for each of Kb sets; first restoring the Kb sets of distributed information based on the Ka mixed images that are obtained; and second restoring the secret information using the Kb sets of distributed information that have been restored.

Method and apparatus for distributed management of files

Sizes of buffer pools for temporarily storing data from a database are dynamically optimized. A database server includes: a dividing ratio setting unit for setting a dividing ratio for dividing a memory area of a predetermined size into a plurality of buffer pools used selectively according to a type of transaction or type of data; a response time measuring unit for measuring response time of transactions; a response time computation formula generating unit for calculating parameters of a response time calculation formula for calculating the response time based on the size of each of the plurality of buffer pools from the set dividing ratio and the measured response time; and a dividing ratio calculation unit for calculating the dividing ratio that minimizes the response time by using the response time calculation formula; wherein the dividing ratio setting unit sets the dividing ratio calculated by the dividing ratio calculating unit.

Growing community of inventors

Yokohama, Japan

Junya Shimizu