US-20260127329-A1 - METHOD FOR AUTOMATICALLY GENERATING DESIGN SOLUTIONS FOR AN ENGINEERING DESIGN PROJECT

Abstract

One variation of a method includes: accessing a descriptor of a project; extracting a set of language signals from the descriptor; accessing a set of input parameters and a set of output characteristics; querying a language model for a range of values of input parameters exhibited within historical engineering solutions correlated with the set of language signals; accessing a virtual model for the project and a function representing a relationship between the set of input parameters and the set of output characteristics; for each analysis instance in a count of analysis instances, defining a combination of values of input parameters within corresponding ranges and based on the virtual model, the function, and the combination of values, executing the analysis instance to calculate a set of values of output characteristics; and rendering representations of combinations of values of input parameters and sets of values of output characteristics within a user interface.

Inventors

• Laurence Cook
• Joe Griston
• Nick Arini
• Nick Boultbee
• Carlos Rodríguez-Toro
• Glyn Stevens
• Ian MacKenzie
• Sam Phillips

Assignees

• GENERATIVE VISION LIMITED

Dates

Publication Date

20260507

Application Date

20251031

Claims (20)

1. 1 . A method comprising: accessing a textual descriptor, supplied by a user, of an engineering design project; extracting a first set of language signals from the textual descriptor; accessing a set of input parameters and a set of output characteristics for the engineering design project; for each input parameter in the set of input parameters, querying a language model for a characteristic range of values of the input parameter exhibited within historical engineering solutions correlated with the first set of language signals; accessing a first set of functions representing relationships between the set of input parameters and the set of output characteristics; accessing a virtual model representing a design solution for the engineering design project; linking a set of model variables in the virtual model to a subset of input parameters in the set of input parameters; for each discover analysis instance in a first count of discover analysis instances: defining a unique combination of discover values of input parameters within corresponding characteristic ranges of the set of input parameters, the unique combination of discover values comprising: a first discover value of a first input parameter within a first characteristic range of values of the first input parameter; and a second discover value of a second input parameter within a second characteristic range of values of the second input parameter; generating a first model instance of the virtual model by modifying a first model variable in the virtual model according to the first discover value of the first input parameter; based on the first model instance, the first set of functions, and the second discover value of the second input parameter, executing a first discover analysis instance to calculate a set of discover values of the set of output characteristics; and rendering representations of the unique combination of discover values of input parameters and the set of discover values of the set of output characteristics within a user interface.
2. 2 . The method of claim 1 : wherein querying the language model for a characteristic range of values of an input parameter for each input parameter in the set of input parameters comprises: querying the language model for a first characteristic range of values of the first input parameter exhibited within historical engineering solutions correlated with the first set of language signals, the first input parameter comprising a geometric variable; and querying the language model for a second characteristic scalar value of the second input parameter exhibited within historical engineering solutions correlated with the first set of language signals, the second input parameter comprising a contextual variable; wherein linking the set of model variables in the virtual model to the subset of input parameters in the set of input parameters comprises linking the first input parameter to the first model variable in the set of model variables in the virtual model; and wherein defining a unique combination of discover values of input parameters for each discover analysis instance in the first count of discover analysis instances comprises: for a first discover analysis instance in the first count of discover analysis instances: defining a first unique combination of discover values of input parameters comprising: the first discover value of the first input parameter comprising the geometric variable; and the second discover value of the second input parameter comprising the second characteristic scalar value.
3. 3 . The method of claim 1 : wherein querying the language model for the characteristic range of values of the input parameter for each input parameter in the set of input parameters comprises: querying the language model for a first characteristic range of values of the first input parameter exhibited within historical engineering solutions correlated with the first set of language signals, the first input parameter comprising a geometric variable; and querying the language model for a second characteristic range of values of the second input parameter exhibited within historical engineering solutions correlated with the first set of language signals, the second input parameter comprising a contextual variable; wherein linking the set of model variables in the virtual model to the subset of input parameters in the set of input parameters comprises linking the first input parameter to a first model variable in the set of model variables in the virtual model; wherein defining the unique combination of discover values of input parameters for each discover analysis instance in the first count of discover analysis instances comprises: for a first discover analysis instance in the first count of discover analysis instances: defining a first unique combination of discover values of input parameters comprising: a third discover value of the first input parameter within the first characteristic range of values of the first input parameter; and a fourth discover value of the second input parameter within the second characteristic range of values of the second input parameter; wherein rendering representations of the unique combination of discover values of input parameters and the set of discover values of the set of output characteristics within the user interface for each discover analysis instance in the first count of discover analysis instances comprises: for the first discover analysis instance: generating a first graphical representation of the first model instance; and rendering the first graphical representation of the first model instance and the first set of discover values of the set of output characteristics within the user interface.
4. 4 . The method of claim 1 , further comprising: in response to execution of the first count of discover analysis instances, receiving a first target range of values of the first input parameter narrowed from the first characteristic range of values; for each explore analysis instance in a second count of explore analysis instances: defining a unique combination of explore values of input parameters comprising an explore value of the first input parameter within the first target range of values of the first input parameter; based on the virtual model, the first set of composite functions, and the unique combination of explore values of input parameters, executing the explore analysis instance to calculate a set of explore values of the set of output characteristics; and rendering representations of the unique combination of explore values of input parameters and the set of explore values of the set of output characteristics within the user interface.
5. 5 . The method of claim 4 , further comprising: receiving a resource limit from the user; estimating a resource allocation for each discover analysis instance; and calculating the first count of discover analysis instances based on the resource limit and the resource allocation; and accessing the second count of explore analysis instances less than the first count of discover analysis instances.
6. 6 . The method of claim 1 : wherein accessing the set of input parameters and the set of output characteristics for the engineering design project comprises defining a first output characteristic in the set of output characteristics based on the first set of language signals; further comprising defining a first target range of values of the first output characteristic based on an input entered by the user into the user interface; wherein defining the unique combination of discover values of input parameters and executing the discover analysis instance for each discover analysis instance in the first count of discover analysis instances comprises: for a first discover analysis instance in the first count of discover analysis instances: defining a first unique combination of discover values of input parameters within corresponding characteristic ranges of the set of input parameters; and based on the virtual model, the first set of functions, and the first unique combination of discover values of input parameters, executing the first discover analysis instance to calculate a first discover value of the first output characteristic; and characterizing a first difference between the first discover value of the first output characteristic and the first target range of values of the first output characteristic; and for a second discover analysis instance in the first count of discover analysis instances: defining a second unique combination of discover values of input parameters within corresponding characteristic ranges of the set of input parameters and predicted to yield a second discover value of the second output characteristic differing from the first target range by less than the first difference; and based on the virtual model and the second unique combination of discover values of input parameters, executing the second discover analysis instance to calculate the second discover value of the second output characteristic.
7. 7 . The method of claim 1 , wherein rendering representations of a unique combination of discover values of input parameters and a set of discover values of the set of output characteristics within the user interface for each discover analysis instance in the first count of discover analysis instances comprises: receiving selection of the first input parameter, in the set of input parameters, from the user; receiving selection of a first output characteristic, in the set of output characteristics, from the user; and rendering a first three-dimensional plot depicting groups of values of the first input parameter and the first output characteristic corresponding to each discover analysis instance in the first count of discover analysis instances.
8. 8 . The method of claim 7 : wherein querying the language model for the characteristic range of values of the input parameter for each input parameter in the set of input parameters comprises: querying the language model for a first characteristic range of values of the first input parameter exhibited within historical engineering solutions correlated with the first set of language signals; and querying the language model for a second characteristic range of values of the second input parameter exhibited within historical engineering solutions correlated with the first set of language signals; and further comprising: prompting the user to isolate a three-dimensional region of interest, within the first three-dimensional plot, containing a subset of groups of values of the first input parameter and the first output characteristic corresponding to discover analysis instances in the first count of discover analysis instances; defining a first target range of values of the first input parameter, narrower than the first characteristic range of values, bounded by the three-dimensional region of interest within the first three-dimensional plot; accessing a second count of explore analysis instances for the engineering design project; for each explore analysis instance in the second count of explore analysis instances: defining a unique combination of explore values of input parameters comprising a first explore value of the first input parameter within the first target range of values of the first input parameter; and based on the virtual model, the first set of functions, and the unique combination of explore values of input parameters, executing the explore analysis instance to calculate a set of explore values of the set of output characteristics; and rendering a second three-dimensional plot depicting groups of values of the first input parameter and the first output characteristic corresponding to each explore analysis instance in the second count of explore analysis instances.
9. 9 . The method of claim 7 : wherein receiving selection of the first output characteristic from the user comprises identifying the first output characteristic as a primary output characteristic of interest to the user based on the first set of language signals; and wherein receiving selection of the first input parameter from the user comprises: deriving a set of correlations between the first output characteristic and each input parameter in the set of input parameters; in response to a first correlation between the first output characteristic and the first input parameter exceeding a threshold correlation: generating a prompt representing the first output characteristic relative to the first input parameter; and serving the prompt to the user via the user interface; and receiving selection of the first input parameter from the user responsive to the prompt.
10. 10 . The method of claim 1 , wherein accessing the set of input parameters and the set of output characteristics comprises: querying the language model for output characteristics considered in design of historical engineering solutions correlated with the first set of language signals; presenting a corpus of output characteristics and corresponding natural language descriptions, received from the language model, within the user interface; and prompting the user to select output characteristics, from the corpus of output characteristics, via the user interface; and receiving selection of the set of output characteristics, from the corpus of output characteristics, via the user input at the user interface.
11. 11 . The method of claim 1 , wherein accessing the set of input parameters and the set of output characteristics comprises: deriving a second set of language signals based on the set of model variables; querying the language model for input parameters considered in design of historical engineering solutions correlated with the first set of language signals and the second set of language signals; presenting a corpus of input parameters and corresponding natural language descriptions, received from the language model, within the user interface; and prompting the user to confirm the set of input parameters, from the corpus of input parameters, via the user interface.
12. 12 . The method of claim 1 : wherein accessing the set of input parameters and the set of output characteristics for the engineering design project comprises: receiving a first natural language descriptor of the first input parameter from the language model; and receiving a second natural language descriptor of the second output characteristic from the language model; and wherein accessing the first set of functions representing relationships between the set of input parameters and the set of output characteristics comprises: querying a function library for: a first function defining: a first independent variable associated with a first set of language concepts congruent with the first natural language descriptor of the first input parameter; and a first dependent variable; and a second function defining: second independent variable; and a second dependent variable associated with language concepts congruent with the second natural language descriptor of the second output characteristic; and based on the first dependent variable and the second independent variable: compiling the first function and the second function to form a first composite function, in the first set of functions; associating the first independent variable with the first input parameter; and associating the second dependent variable with the second output characteristic.
13. 13 . The method of claim 1 : wherein accessing the set of input parameters and the set of output characteristics for the engineering design project comprises receiving a first natural language descriptor of the first input parameter from the language model; and wherein linking the set of model variables in the virtual model to the subset of input parameters in the set of input parameters comprises: detecting a first model variable, in the set of model variables, in the virtual model; extracting a first language concept from the first model variable in the virtual model; calculating a correspondence between the first language concept and the first natural language descriptor; and in response to the correspondence exceeding a threshold correspondence, linking the first input parameter to the first model variable.
14. 14 . The method of claim 1 : wherein accessing the set of input parameters and the set of output characteristics for the engineering design project comprises receiving a first natural language descriptor of the first input parameter from the language model; and wherein linking the set of model variables in the virtual model to the subset of input parameters in the set of input parameters comprises: detecting a first model variable, in the set of model variables, in the virtual model; defining a first test value and a second test value, different from the first test value, for the first model variable; generating a first image of the virtual model rendered according to the first model variable defined by the first test value; generating a second image of the virtual model rendered according to the first model variable defined by the second test value; querying a language model for a second natural language descriptor of a difference between the first image and the second image; calculating a first correspondence between the first natural language descriptor and the second natural language descriptor; and in response to the first correspondence exceeding a threshold, linking the first input parameter to the first model variable.
15. 15 . The method of claim 1 : further comprising extracting a second set of language signals, descriptive of the design solution, from symbols representing the set of model variables in the virtual model; and wherein querying the language model for the characteristic range of values of the input parameter for each input parameter in the set of input parameters comprises, for each input parameter in the set of input parameters: querying the language model for the characteristic range of values of the input parameter exhibited within historical engineering solutions correlated with: the first set of language signals; and the second set of language signals.
16. 16 . A method comprising, at a remote computer system: accessing a virtual model representing a design solution for an engineering design project; deriving a first set of language signals representing the engineering design project based on the virtual model; accessing a set of input parameters and a set of output characteristics for the engineering design project; for each input parameter in the set of input parameters: querying a language model for a characteristic range of values of the input parameter exhibited within historical engineering solutions correlated with the first set of language signals; accessing a first set of functions representing relationships between the set of input parameters and the set of output characteristics; linking a set of model variables in the virtual model to a subset of input parameters in the set of input parameters; and for each discover analysis instance in a first set of discover analysis instances: defining a unique combination of discover values of input parameters within corresponding characteristic ranges of the set of input parameters; based on the virtual model, the first set of functions, and the unique combination of discover values of input parameters, executing the discover analysis instance to calculate a set of discover values of the set of output characteristics; and rendering representations of the unique combination of discover values of input parameters and the set of discover values of the set of output characteristics within a user interface.
17. 17 . The method of claim 16 : wherein accessing the set of input parameters and the set of output characteristics for the engineering design project comprises: querying the language model for the set of input parameters and the set of output characteristics considered in design of historical engineering solutions correlated with the first set of language signals; receiving a first natural language descriptor of the first input parameter, in the set of input parameters, from the language model; and receiving a second natural language descriptor of a second output characteristic, in the set of output characteristics, from the language model; wherein accessing the first set of functions comprises: accessing a first function defining: a first independent variable associated with a first set of language concepts congruent with the first natural language descriptor of the first input parameter; and a first dependent variable; and accessing a second function defining: a second independent variable; and a second dependent variable associated with language concepts congruent with the second natural language descriptor of the second output characteristic; and based on congruence between the first dependent variable and the second independent variable: compiling the first function and the second function into a first composite function, in the first set of functions; associating the first independent variable with the first input parameter; and associating the second dependent variable with the second output characteristic; further comprising: accessing a textual descriptor of the engineering design project; and extracting a second set of language signals from the textual descriptor; wherein querying the language model for the characteristic range of values of the input parameter for each input parameter in the set of input parameters comprises, for each input parameter in the set of input parameters: querying the language model for the characteristic range of values of the input parameter exhibited within historical engineering solutions correlated with: the first set of language signals; and the second set of language signals.
18. 18 . The method of claim 16 : further comprising extracting a set of symbols, descriptive of the design solution, from the set of model variables in the virtual model; and wherein querying the language model for the characteristic range of values of the input parameter for each input parameter in the set of input parameters comprises, for each input parameter in the set of input parameters: querying the language model for the characteristic range of values of the input parameter exhibited within historical engineering solutions correlated with: the first set of language signals; and the set of symbols.
19. 19 . A method comprising, at a remote computer system: accessing a textual descriptor of an engineering design project; extracting a first set of language signals from the textual descriptor; accessing a virtual model representing design solutions for the engineering design project; deriving a second set of language signals based on a set of model variables within the virtual model; accessing a corpus of input parameters based on the first set of language signals and the second set of language signals; presenting the corpus of input parameters within a user interface; based on selection of a set of input parameters from the corpus of input parameters via the user interface, querying a language model for characteristic ranges of values of input parameters in the set of input parameters; accessing a function representing relationships between the set of input parameters and a set of output characteristics; for a first analysis instance: defining a first combination of values of input parameters within corresponding characteristic ranges; and based on the virtual model, the function, and the first combination of values of input parameters, executing the first analysis instance to calculate a first set of values of the set of output characteristics; and rendering a first representation of the first combination of values of input parameters and the first set of values of the set of output characteristics, from the first analysis instance, within the user interface.
20. 20 . The method of claim 19 , further comprising: for a second analysis instance: defining a second combination of values of input parameters within corresponding characteristic ranges; and based on the virtual model, the function, and the second combination of values of input parameters, executing the second analysis instance to calculate a second set of values of the set of output characteristics; and rendering a second representation of the second combination of values of input parameters and the second set of values of the set of output characteristics, from the second analysis instance, adjacent the first representation within the user interface.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS This Application is a continuation of U.S. patent application Ser. No. 18/965,863, filed on 02-DEC-2024, which claims the benefit of U.S. Provisional Application No. 63/604,766, filed on 30-NOV-2023, and 63/640,804, filed on 30-APR-2024, each of which is incorporated in its entirety by this reference. TECHNICAL FIELD This invention relates generally to the field of product design and more specifically to a new and useful method for automatically generating design solutions in the field of product design. BRIEF DESCRIPTION OF THE FIGURES FIGS. 1A, 1B, and 1C are flowchart representations of a method; FIG. 2 is a flowchart representation of one variation of the method; FIG. 3 is a flowchart representation of one variation of the method; FIG. 4 is a flowchart representation of one variation of the method; FIGS. 5A and 5B are flowchart representations of one variation of the method; and FIG. 6 is a flowchart representation of one variation of the method. DESCRIPTION OF THE EMBODIMENTS The following description of embodiments of the invention is not intended to limit the invention to these embodiments but rather to enable a person skilled in the art to make and use this invention. Variations, configurations, implementations, example implementations, and examples described herein are optional and are not exclusive to the variations, configurations, implementations, example implementations, and examples they describe. The invention described herein can include any and all permutations of these variations, configurations, implementations, example implementations, and examples. 1. Method As shown in FIGS. 1A, 1B, and 1C a method includes: accessing a textual descriptor, supplied by a user, of an engineering design project in Block S110; extracting a first set of language signals from the textual descriptor in Block S112; accessing a set of input parameters and a set of output characteristics for the engineering design project in Block S120; and, for each input parameter in the set of input parameters, querying a language model for a characteristic range of values of the input parameter exhibited within historical engineering solutions correlated with the first set of language signals in Block S130. The method S100 further includes: accessing a first set of composite functions representing relationships between the set of input parameters and the set of output characteristics in Block S140; accessing a virtual model representing a design solution for the engineering design project in Block S150; linking a set of model variables in the virtual model to a subset of input parameters in the set of input parameters in Block S152; and accessing a first count of discover analysis instances for the engineering design project in Block S160. The method S100 further includes, for each discover analysis instance in the first count of discover analysis instances: defining a unique combination of discover values of input parameters within corresponding characteristic ranges of the set of input parameters in Block S162; based on the virtual model, the first set of composite functions, and the unique combination of discover values of input parameters, executing the discover analysis instance to calculate a set of discover values of the set of output characteristics in Block S164; and rendering representations of the unique combination of discover values of input parameters and the set of discover values of the set of output characteristics within a user interface in Block S166. 1.1 Variation: Virtual Model Influences Range Selection One variation of the method S100, includes: accessing a virtual model representing a design solution for an engineering design project in Block S150; deriving a first set of language signals representing the engineering design project based on the virtual model in Block S112; accessing a set of input parameters and a set of output characteristics for the engineering design project in Block S120; for each input parameter in the set of input parameters, querying a language model for a characteristic range of values of the input parameter exhibited within historical engineering solutions correlated with the first set of language signals in Block S130; accessing a first set of composite functions representing relationships between the set of input parameters and the set of output characteristics in Block S140; linking a set of model variables in the virtual model to a subset of input parameters in the set of input parameters in Block S152; and accessing a first count of discover analysis instances for the engineering design project in Block S160. This variation of the method S100 further includes, for each discover analysis instance in the first count of discover analysis instances: defining a unique combination of discover values of input parameters within corresponding characteristic ranges of the set of input parameters in Block S162; based on the virtual model, the