Method and system for designing formulated products

Patent metadata

Field	Value
Publication number	US-11488226-B2
Application number	US-202016909845-A
Country	US
Kind code	B2
Filing date	Jun 23, 2020
Priority date	Jun 24, 2019
Publication date	Nov 1, 2022
Grant date	Nov 1, 2022

Abstract

This disclosure relates generally to method and system for designing the formulated products. Conventional techniques for designing the formulated products, meeting final functional properties, are limited. Further, understanding user requirements and active incorporation of the user requirements during design phase is quite challenging. The present disclosure herein provides method and system that solve the technical problem of extracting the functional requirement by establishing continuous conversation with the user. An optimal prediction function for each functional requirement is determined by using a plurality of prediction models. An optimization technique along with an objective function is employed to determine optimized solutions comprising list of ingredients, possible concentration level of each ingredient, the process parameters, and the operating parameters for obtaining the desired formulation based on the user requirement.

First claim

What is claimed is:

1. A processor-implemented method comprising the steps of:
receiving, via one or more hardware processors, an intent associated with a desired formulation of a target formulated product, from a user;
extracting, via the one or more hardware, processors, one or more functional parameters and one or more performance indicators of the desired formulation, based on the received intent;
obtaining, via the one or more hardware processors, a quantitative value for each functional parameter of the one or more functional parameters, using a conversion look-up table;
identifying, via the one or more hardware processors, one or more key input parameters associated with the one or more functional parameters of the desired formulation, based on the one or more performance indicators, wherein the one or more key input parameters include (i) one or more process parameters, (ii) one or more operating condition parameters, (iii) one or more ingredients, and (iv) one or more parameters associated with the one or more ingredients;
determining, via the one or more hardware processors, an optimal prediction model for each functional parameter of the one or more functional parameters, to obtain an optimal prediction function for the associated functional parameter; and
determining, via the one or more hardware processors, an optimal solution dataset of the one or more key input parameters associated with the one or more functional parameters for the desired formulation, using an optimization technique, based on an objective function, and one or more constraints comprising (i) the quantitative value for each functional parameter of the one or more functional parameters, (ii) lower bound values and upper bound values of the one or more key input parameters, and (iii) the one or more performance indicators, wherein the objective function is defined as a weighted sum of the optimal prediction function for each functional parameter of the one or more functional parameters.

2. The method as claimed in claim 1 further comprising displaying, via the one or more hardware processors, the optimal solution dataset of the one or more key input parameters associated with the one or more functional parameters of the desired formulation, on visualization tools.

3. The method as claimed in claim 1 , wherein the one or more functional parameters and the one or more performance indicators of the desired formulation are extracted based on the received intent, by establishing continuous conversation with the user, using one or more conversational agents that are trained with intent-action mechanism based training dataset.

4. The method as claimed in claim 1 , wherein determining the optimal prediction model for each functional parameter of the one or more functional parameters, comprises:
obtaining an input dataset associated with the one or more key input parameters and the one or more functional parameters of the desired formulation, wherein the input dataset comprises data elements for the one or more key input parameters and the one or more functional parameters;
extracting a sub-input dataset of each functional parameter, from the input dataset, wherein the sub-input dataset comprises the data elements for the one or more key input parameters and the associated functional parameter;
pre-processing the sub-input dataset to obtain a pre-processed dataset of each functional parameter, wherein the pre-processed dataset comprises pre-processed data elements for the one or more key input parameters and the associated functional parameter;
dividing the pre-processed dataset of each functional parameter into a training dataset and a testing dataset, based on a predefined ratio;
scaling the training dataset of each functional parameter to obtain a scaled training dataset;
scaling the testing dataset of each functional parameter to obtain a scaled testing dataset;
generating one or more prediction models of each functional parameter, using the scaled training dataset;
validating the one or more generated prediction models of each functional parameter using the scaled testing dataset;
determining the optimal prediction model for each functional parameter out of the one or more generated prediction models, based on the validation; and
obtaining the optimal prediction function for the associated functional parameter, from the optimal prediction model.

5. The method as claimed in claim 4 , wherein preprocessing the sub-input dataset to obtain the pre-processed dataset of each functional parameter, comprises imputing missing data, outlier removal, and high correlation coefficient data removal.

6. A system comprising:
a memory storing instructions;
one or more Input/Output (I/O) interfaces; and
one or more hardware processors coupled to the memory via the one or more I/O interfaces, wherein the one or more hardware processors are configured by the instructions to:
receive an intent associated with a desired formulation of a target formulated product, from a user;
extract one or more functional parameters and one or more performance indicators of the desired formulation, based on the received intent;
obtain a quantitative value for each functional parameter of the on or more functional parameters, using a conversion look-up table;
identify one or more key input parameters associated with the one or more functional parameters of the desired formulation, based on the one or more performance indicators, wherein the one or more key input parameters include (i) one or more process parameters, (ii) one or more operating condition parameters, (iii) one or more ingredients, and (iv) one or more parameters associated with the one or more ingredients;
determine an optimal prediction model for each functional parameter of the one or more functional parameters, to obtain an optimal prediction function for the associated functional parameter; and
determine an optimal solution dataset of the one or more key input parameters associated with the one or more functional parameters for the desired formulation, using an optimization technique, based on an objective function, and one or more constraints comprising (i) the quantitative value for each functional parameter of the one or more functional parameters, (ii) lower bound values and upper bound values of the one or more key input parameters, and (iii) the one or more performance indicators, wherein the objective function is defined as a weighted sum of the optimal prediction function for each functional parameter of the one or more functional parameters.

7. The system as claimed in claim 6 , wherein the one or more hardware processors are further configured to display the optimal solution dataset of the one or more key input parameters associated with the one or more functional parameters of the desired formulation, on visualization tools.

8. The system as claimed in claim 6 , wherein the one or more hardware processors are further configured to extract the one or more functional parameters and the one or more performance indicators of the desired formulation, by establishing continuous conversation with the user, based on the received intent, using one or more conversational agents that are trained with intent-action mechanism based training dataset.

9. The system as claimed in claim 6 , wherein the one or more hardware processors are further configured to determine the optimal prediction model for each functional parameter of the one or more functional parameters, by;
obtaining an input dataset associated with the one or more key input parameters and the one or more functional parameters of the desired formulation, wherein the input dataset comprises data elements for the one or more key input parameters and the

Assignees

• Tata Consultancy Services Ltd
• Tata Consultancy Ltd Services

Inventors

• Daware Santosh Vasant
• Chako Rinu
• Jain Deepak Shyamsunder
• Kausley Shankar Balajirao
• Gupta Shally
• Bandi Pallavi
• Kumar Dharmendr
• Rai Beena
• Bhowmik Amit
• Singh Umesh
• Malhotra Chetan Premkumar
• Basavarsu Purushottham Gautham

Classifications

• G06N20/20 Primary

  Ensemble learning · CPC title

• G06Q30/0631 Primary

  Recommending goods or services · CPC title

• G06F18/214

  Generating training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• G06N5/00

  Computing arrangements using knowledge-based models · CPC title

• Y10S707/99931

  Database or file accessing · CPC title