Minimizing uncertainty envelopes in trajectories of evolving ensemble members

Patent metadata

Field	Value
Publication number	US-10929578-B2
Application number	US-201715853192-A
Country	US
Kind code	B2
Filing date	Dec 22, 2017
Priority date	Apr 23, 2014
Publication date	Feb 23, 2021
Grant date	Feb 23, 2021

Abstract

A mechanism is provided in a data processing system for minimizing uncertainty envelopes in trajectories of evolving ensemble members. The mechanism generates a trajectory forecast of each member object of an ensemble based on an initial state-space and a model for predicting trajectories of the member objects to generate a plurality of trajectory forecasts. Each of the plurality of trajectory forecasts has an individual uncertainty envelope. The mechanism applies a classification algorithm on the plurality of trajectory forecasts to identify at least one group of member objects having similar trajectory forecasts, generates a reduced ensemble of member objects including the identified group of member objects, and reconfigures the state-space and the model for predicting trajectories. The mechanism generates an updated trajectory forecast of each member object of the reduced ensemble based on the reconfigured state-space and the reconfigured model for predicting trajectories of the member objects.

First claim

What is claimed is:

1. A method, in a data processing system comprising at least one processor and at least one memory, the at least one memory comprising instructions executed by the at least one processor to cause the at least one processor to implement a simulation system for minimizing uncertainty envelopes in trajectories of evolving ensemble members, the method comprising:
receiving, by the simulation system from a user, a set of defining parameters for selecting member objects of an ensemble;
identifying, by the simulation system, the ensemble of member objects satisfying the defining parameters;
measuring initial states and parameter values for the ensemble of member objects based on an initial state-space;
generating, by a predictor executing within the simulation system, a trajectory forecast of each member object of the ensemble based on the initial state-space, the initial states and parameter values, and a model for predicting trajectories of the member objects to generate a plurality of trajectory forecasts, wherein each of the plurality of trajectory forecasts has an individual uncertainty envelope and wherein the ensemble has an ensemble uncertainty envelope;
applying, by a clustering system executing within the simulation system, a classification algorithm on the plurality of trajectory forecasts to identify at least one group of member objects having similar trajectory forecasts;
generating, by an envelope calculator executing within the simulation system, a reduced ensemble of member objects including the identified group of member objects;
performing at least one targeted correction to the reduced ensemble of member objects that is expected to cause the reduced ensemble to behave similarly;
reconfiguring, by the simulation system, the state-space and the model for predicting trajectories for the reduced ensemble of member objects based on the at least one targeted correction;
measuring updated states and parameter values for the reduced ensemble based on the reconfigured state-space; and
generating, by the predictor executing within the simulation system, an updated trajectory forecast of each member object of the reduced ensemble based on the reconfigured state-space, the updated states and parameter values, and the reconfigured model for predicting trajectories of the member objects.

2. The method of claim 1 , wherein each ensemble member has measurable properties that serve as state variables to define its location in the state-space at any given time and intrinsic parameters that determine changes in location from one time to another in the state-space and wherein the model predicts the future location of each ensemble member given its current measurable properties and its intrinsic parameters.

3. The method of claim 1 , further comprising:
calculating a predicted uncertainty envelope of trajectories of the reduced ensemble.

4. The method of claim 3 , wherein calculating the predicted uncertainty envelope of trajectories comprises applying a minimax estimate function to the reduced ensemble.

5. The method of claim 1 , wherein applying the classification algorithm on the plurality of trajectory forecasts comprises identifying a plurality of sub-groups of member objects having similar trajectory forecasts.

6. The method of claim 5 , further comprising applying a minimax estimate function to each sub-group of member objects to predict a respective uncertainty envelope of trajectories.

7. The method of claim 1 , wherein the classification algorithm comprises a k-means clustering algorithm.

8. The method of claim 1 , wherein each member object within the ensemble of member objects comprises a device being manufactured using a staged process starting from initial components and leading to a final product, wherein the initial state-space comprises a set of test measurements during assembly, and wherein the at least one targeted correction comprises an adjustment to at least one component.

9. A computer program product comprising a computer readable storage medium having a computer readable program stored therein, wherein the computer readable program, when executed on a computing device, causes the computing device to implement a simulation system for minimizing uncertainty envelopes in trajectories of evolving ensemble members, the computer readable program causing the computing device to:
receive, by the simulation system from a user, a set of defining parameters for selecting member objects of an ensemble;
identify, by the simulation system, the ensemble of member objects satisfying the defining parameters;
measure initial states and parameter values for the ensemble of member objects based on an initial state-space;
generate, by a predictor executing within the simulation system, a trajectory forecast of each member object of the ensemble based on the initial state-space, the initial states and parameter values, and a model for predicting trajectories of the member objects to generate a plurality of trajectory forecasts, wherein each of the plurality of trajectory forecasts has an individual uncertainty envelope and wherein the ensemble has an ensemble uncertainty envelope;
apply, by a clustering system executing within the simulation system, a classification algorithm on the plurality of trajectory forecasts to identify at least one group of member objects having similar trajectory forecasts;
generate, by an envelope calculator executing within the simulation system, a reduced ensemble of member objects including the identified group of member objects;
perform at least one targeted correction on the reduced ensemble of member objects that is expected to cause the reduced ensemble to behave similarly;
reconfigure, by the simulation system, the state-space and the model for predicting trajectories for the reduced ensemble of member objects based on at least one targeted correction performed on the reduced ensemble of member objects;
measure updated states and parameter values for the reduced ensemble based on the reconfigured state-space; and
generate, by the predictor executing within the simulation system, an updated trajectory forecast of each member object of the reduced ensemble based on the reconfigured state-space, the updated states and parameter values, and the reconfigured model for predicting trajectories of the member objects.

10. The computer program product of claim 9 , wherein the computer readable program further causes the computing device to:
calculate a predicted uncertainty envelope of trajectories of the reduced ensemble.

11. The computer program product of claim 10 , wherein calculating the predicted uncertainty envelope of trajectories comprises applying a minimax estimate function to the reduced ensemble.

12. The computer program product of claim 9 , wherein the classification algorithm comprises a k-means clustering algorithm.

13. The computer program product of claim 9 , wherein each ensemble member has measurable properties that serve as state variables to define its location in the state-space at any given time and intrinsic parameters that determine changes in location from one time to another in the state-space and wherein the model predicts the future location of each ensemble member given its current measurable properties and its intrinsic parameters.

14. The computer program product of claim 9 , wherein each member object within the ensemble of member objects comprises a device being manufactured using a staged process starting from initial components and leading to a final product, wherein the initial state-space comprises a set of test measurements during assembly, a

Assignees

• IBM

Inventors

• O'Donncha Fearghal
• Ragnoli Emanuele
• Suits Frank
• Zhuk Sergiy

Classifications

• G06F17/18 Primary

  for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)} · CPC title

• G06F30/20 Primary

  Design optimisation, verification or simulation (optimisation, verification or simulation of circuit designs G06F30/30) · CPC title