House of quality for federated manufacturing apps

Patent metadata

Field	Value
Publication number	US-10990360-B2
Application number	US-201716340851-A
Country	US
Kind code	B2
Filing date	Sep 28, 2017
Priority date	Oct 24, 2016
Publication date	Apr 27, 2021
Grant date	Apr 27, 2021

Abstract

A computer-implemented method of identifying federations of manufacturing apps using House of Quality (HoQ) analysis includes receiving user request information comprising a plurality of desired functions and an importance value for each of the plurality of desired functions and generating a correlation matrix between the plurality of apps, wherein each element of the correlation matrix indicates a degree of correlation between a pair of apps. A relationship matrix is generated which comprises numerical values indicating relationships between a plurality of apps and the plurality of desired functions, wherein each column of relationship matrix is associated with an app and each row is associated with a desired function. Next, a weighted matrix is created by multiplying each row in the relationship matrix by the importance value corresponding to the desired function associated with the row. The numerical values in each column of the weighted matrix are aggregated to determine a total value for each of the plurality of apps. A federation of apps is identified based on the correlation matrix and the total values determined for the plurality of apps.

First claim

We claim:

1. A computer-implemented method of identifying federations of manufacturing apps using house of quality (HoQ) analysis, the method comprising:
receiving user request information comprising a plurality of desired functions and an importance value for each of the plurality of desired functions;
generating a correlation matrix between the plurality of apps, wherein each element of the correlation matrix indicates a degree of correlation between a pair of apps, wherein the correlation matrix is generated at least in part by:
for each app, performing a semantic analysis to determine a similarity measure for each of the other apps indicating a degree of similarity with respect to source code information; and
populating elements of the correlation matrix between two apps with the similarity measure corresponding to the two apps;

generating a relationship matrix comprising numerical values indicating relationships between a plurality of apps and the plurality of desired functions, wherein each column of the relationship matrix is associated with an app and each row is associated with a desired function;
creating a weighted matrix by multiplying each row in the relationship matrix by the importance value corresponding to the desired function associated with the row;
aggregating the numerical values in each column of the weighted matrix to determine a total value for each of the plurality of apps; and
identifying a federation of apps based on the correlation matrix and the total values determined for the plurality of apps.

2. The method of claim 1 , wherein the source code information comprises one or more of class name, function name, variable name, or output values.

3. The method of claim 1 , wherein the source code information comprises one or more of function block input parameters and function block output parameters.

4. The method of claim 1 , wherein the method further comprises:
analyzing records of app installations to identify combinations of apps previously installed by one or more users,
wherein the correlation matrix is generated, at least in part, by populating elements of the correlation matrix between the pair of apps with a similarity measure indicating that the pair of apps are both included in two or more of the combinations of apps.

5. The method of claim 1 , wherein the method further comprises:
analyzing records of app usage to identify combinations of apps previously used by one or more users,
wherein the correlation matrix is generated, at least in part, by populating elements of the correlation matrix between the pair of apps with a similarity measure indicating that the pair of apps are both included in two or more of the combinations of apps.

6. The method of claim 1 , wherein the relationship matrix is generated at least in part by:
applying a machine learning model to each app and the each desired function to determine a numerical value indicating a degree of relatedness between the app and the desired function; and
for each app and each desired function, populating an element of the relationship matrix corresponding to the app and the desired function with the numerical value indicating the degree of relatedness between the app and the desired function.

7. The method of claim 1 , further comprising:
generating a HoQ diagram comprising a visualization of:
the plurality of desired functions,
the importance value for each of the plurality of functions,
the correlation matrix,
the relationship matrix, and
the total values determined for the plurality of apps; and
presenting the HoQ diagram on at least one display.

8. The method of claim 7 , further comprising:
determining a relative weight for each of the total values by dividing each total value by a summation of all of the total values,
wherein the HoQ diagram further comprises a visualization of the relative weight for each of the total values.

9. The method of claim 1 , wherein receiving the user request information comprises:
receiving a requirements document from a user;
applying a natural language processing model to the requirements document to extract the plurality of desired functions and the importance value for each of the plurality of desired functions.

10. The method of claim 9 , further comprising:
automatically installing each app included in the federation of apps on a user computing system.

11. A computer-implemented method of identifying recommendations of manufacturing apps using house of quality (HoQ) analysis, the method comprising:
receiving, from a user, user request information comprising a plurality of desired functions and an importance value for each of the plurality of desired functions;
generating a HoQ diagram comprising (a) a relationship matrix comprising numerical values indicating relationships between the plurality of apps and the plurality of desired functions and (b) a correlation matrix between pairs of the plurality of apps, wherein the correlation matrix is generated at least in part by:
for each app, performing, a semantic analysis to determine a similarity measure for each of the other apps indicating a degree of similarity with respect to source code information; and
populating elements of the correlation matrix between two apps with the similarity measure corresponding to the two apps;

using the HoQ diagram to identify one or more app recommendations providing the plurality of desired functions; and
presenting the one or more app recommendations to the user.

12. The method of claim 11 , wherein the source code information comprises one or more of class name, function name, or variable name.

13. The method of claim 11 , wherein the source code information comprises one or more of function block input parameters and function block output parameters.

14. The method of claim 11 , wherein the method further comprises:
analyzing records of app installations to identify combinations of apps previously installed by one or more users,
wherein the correlation matrix is determined, at least in part, by populating elements of the correlation matrix between two apps with a similarity measure indicating that the two apps are both included in two or more of the combinations of apps.

15. The method of claim 11 , wherein the method further comprises:
analyzing records of app usage to identify combinations of apps previously used by one or more users,
wherein the correlation matrix is determined, at least in part, by populating elements of the correlation matrix between two apps with a similarity measure indicating that the two apps are both included in two or more of the combinations of apps.

16. The method of claim 11 , wherein receiving the user request information comprises:
receiving a requirements document from a user;
applying a natural language processing model to the requirements document to extract the plurality of desired functions and the importance value for each of the plurality of desired functions.

17. The method of claim 16 , further comprising:
automatically installing each app included in the app recommendations on a user computing system.

18. A computer-implemented method of identifying recommendations of manufacturing apps using house of quality (HoQ) analysis, the method comprising:
receiving, from a plurality of app developers, a plurality of apps;
analyzing each of the plurality of apps to determine design features of the plurality of apps;
receiving, from a user, a plurality of desired functions;
generating a HoQ diagram comprising (a) a relationship matrix co

Assignees

• Siemens Ag
• Georgia Tech Res Inst

Inventors

• Gruenewald Thomas
• Wang Lingyun
• Rosca Justinian
• Kurfess Thomas
• Locks Stephanie
• Chen Aoyu
• Han Dongmin
• Huang Nuodi
• Lynn Roby

Classifications

• G06F8/35 Primary

  model driven · CPC title

• G06F8/20

  Software design · CPC title

• G06N20/00

  Machine learning · CPC title

• G06Q10/06395

  Quality analysis or management · CPC title

• G06F40/30

  Semantic analysis · CPC title