Generative group-based meal planning system and method

• US 10,373,522 B2
• Filed: 01/05/2016
• Issued: 08/06/2019
• Est. Priority Date: 01/05/2016
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Byron, Donna K., DiMascio, Carmine M., Pinel, Florian, Winkler, Timothy P.
• Primary Examiner(s): Yao, Samchuan C
• Assistant Examiner(s): Humphrey, Michael C
• Application Number:
  US14/988,357
• Publication Number:
  US 20170193853A1
• Time in Patent Office: 1,309 Days
• Field of Search: 434127
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G09B 19/0092 Nutrition

Abstract

Embodiments provide a generative group-based meal planning system and method for the creation of candidate meal plans based upon a pre-selected list of ingredients. The meal planning system can create parent meal plans based upon one or more recipes having one or more of the pre-selected ingredients. Child meal plans can be created by the random crossing of the recipes contained in the parent meal plans. The child meal plans can be scored against a genetic algorithm, such as a fitness function, which takes into consideration cost of ingredients, waste, flavor compatibility, preparation time, and ingredient shelf life. The meal planning system can utilize a cognitive system with natural language processing abilities to generate new recipes based off of waste or flavor compatibility. The child meal plans having the highest fitness score can be used as the parent meal plans in the next iteration of analysis. After a pre-determined number of iterations, a candidate meal plan can be output by the system.

Claims

1. A computer implemented method in a data processing system comprising a processor and a memory comprising instructions which are executed by the processor to cause the processor to implement a generative group-based meal planning system comprising a cognitive system having natural language processing capabilities, the method comprising:

receiving, by the meal planning system, a request to generate a candidate meal plan, wherein a seed list including one or more ingredients is designated by the cognitive system based on dish classifications, categories of ingredients, and cuisine classifications;

wherein the one or more ingredients included in the seed list are incorporated into one or more recipes contained in the candidate meal plan, wherein the number of needed recipes to be included in the candidate meal plan is pre-determined;

importing, by the meal planning system, one or more recipes incorporating the one or more identified ingredients;

generating, by the cognitive system, one or more parent meal plans, each parent meal plan containing the pre-determined number of recipes, wherein the parent meal plan recipes are randomly selected from the imported recipes incorporating the one or more identified ingredients;

generating, by the meal planning system, one or more child meal plans, wherein each child meal plan is generated by randomly selecting different parent meal plan recipes and combining the selected parent meal plan recipes;

wherein one or more mutations are strategically inserted by the cognitive system into the one or more child meal plans;

determining, by the meal planning system, a fitness score for the one or more child meal plans through the utilization of a genetic algorithm, wherein the genetic algorithm comprises a fitness function that considers cost of ingredients, shelf life of ingredients, flavor compatibility of ingredients, waste ingredients, and preparation time as factors in determining the fitness score, wherein the cognitive system generates an additional recipe using all the waste ingredients, and the additional recipe is incorporated into the one or more parent meal plans;

wherein the steps of generating one or more parent meal plans, generating one or more child meal plans, and determining the fitness score for the one or more child meal plans are repeated for a pre-determined and finite number of iterations, wherein the one or more child meal plans with the highest fitness scores are used as the next generation of one or more parent meal plans; and

outputting, by the meal planning system, the candidate meal plan having the pre-determined number of recipes, wherein the child meal plan having the highest fitness score is selected to be the candidate meal plan.

2. The method as recited in claim 1, wherein in addition to identifying one or more ingredients to be incorporated, one or more ingredients to be excluded from the one or more recipes contained in the candidate meal plan are further identified.

3. The method as recited in claim 1, wherein the meal planning system is connected to a database containing cost of ingredient information.

4. The method as recited in claim 1, wherein the meal planning system is connected to a database containing shelf life of ingredient information.

5. The method as recited in claim 1, wherein the meal planning system is connected to a database containing flavor compatibility of ingredient information.

6. The method as recited in claim 1, further comprising:

caching, through the meal planning system, waste ingredients in one or more recipes identified during the determination of the fitness score;

creating, through the cognitive system, one or more new recipes based upon the waste ingredients cached; and

incorporating, by the meal planning system, the one or more new recipes based upon the waste ingredients into the next generation of one or more parent meal plans.

7. The method as recited in claim 1, further comprising:

identifying, through the cognitive system, one or more identified ingredients as having flavor compatibility;

creating, through the cognitive system, one or more new recipes based upon the compatible ingredients; and

incorporating, by the meal planning system, the one or more new recipes based upon the compatible ingredients into the first or next generation of one or more parent meal plans.

8. The method as recited in claim 7, wherein the fitness function is configured to award greater weight to recipes having greater cost efficiency of ingredients, thereby causing the cognitive system to create the one or more new recipes prioritizing cost of ingredients over flavor compatibility.

9. The method as recited in claim 1, further comprising the step of:

verifying the one or more recipes in the candidate meal plan against shelf life of ingredient information.

10. The method as recited in claim 9, further comprising the step of:

reorganizing the one or more recipes in the candidate meal plan to maximize the shelf life of the ingredients.

18. The method as recited in claim 1, wherein the one or more mutations comprise one or more of:

altering an ingredient in a particular recipe included in the one or more child meal plans, substituting a first recipe for a second recipe, and altering the amount of ingredients in the particular recipe due to a waste analysis from the fitness function.

11. A non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to:

receive a request to generate a candidate meal plan, wherein a seed list including one or more ingredients is designated by a cognitive system based on dish classifications, categories of ingredients, and cuisine classifications, wherein the one or more ingredients included in the seed list are incorporated into one or more recipes contained in the candidate meal plan, wherein the number of needed recipes to be included in the candidate meal plan is pre-determined;

import one or more recipes incorporating the one or more identified ingredients;

generate, by the cognitive system, one or more parent meal plans each parent meal plan containing the pre-determined number of recipes, wherein the parent meal plan recipes are randomly selected from the identified recipes incorporating the one or more imported ingredients;

generate one or more child meal plans, wherein each child meal plan is generated by randomly selecting different parent meal plan recipes and combining the selected parent meal plan recipes;

wherein one or more mutations are strategically inserted by the cognitive system into the one or more child meal plans;

determine a fitness score for the one or more child meal plans through the utilization of a genetic algorithm, wherein the genetic algorithm comprises a fitness function that considers cost of ingredients, shelf life of ingredients, flavor compatibility of ingredients, waste ingredients, and preparation time as factors in determining the fitness score;

wherein the cognitive system generates an additional recipe using all the waste ingredients, and the additional recipe is incorporated into the one or more parent meal plans; and

output the candidate meal plan having the chosen number of recipes, wherein the child meal plan having the highest fitness score is selected to be the candidate meal plan;

wherein generating one or more parent meal plans, generating one or more child meal plans, and determining the fitness score for the one or more child meal plans are repeated for a pre-determined and finite number of iterations, wherein the one or more child meal plans with the highest fitness scores are used as the next generation of one or more parent meal plans.

12. The computer readable storage medium as recited in claim 11, wherein the processor is further configured to identify one or more ingredients to be excluded from the one or more recipes contained in the candidate meal plan.

13. The computer readable storage medium as recited in claim 11, wherein the processor is connected to a database containing flavor compatibility of ingredient information.

14. The computer readable storage medium as recited in claim 11, wherein the cognitive system has natural language processing capabilities.

15. The computer readable storage medium as recited in claim 14, wherein the processor is configured to cache waste ingredients in one or more recipes identified during the determination of the fitness score;

the cognitive system is configured to create one or more new recipes based upon the waste ingredients cached; and

the processor is further configured to incorporate the one or more new recipes based upon the waste ingredients into the next generation of one or more parent meal plans.

16. The computer readable storage medium as recited in claim 14,

wherein

the cognitive system is configured to:

identify one or more identified ingredients as having flavor compatibility; and

create one or more new recipes based upon the compatible ingredients; and

the processor is further configured to incorporate the one or more new recipes based upon the compatible ingredients into the first or next generation of one or more parent meal plans.

19. The computer readable storage medium as recited in claim 11, wherein the one or more mutations comprise one or more of:

altering an ingredient in a particular recipe included in the one or more child meal plans, substituting a first recipe for a second recipe, and altering the amount of ingredients in the particular recipe due to a waste analysis from the fitness function.

17. A system for generative group-based meal planning, comprising:

a cognitive system having natural language processing capabilities;

a meal planning processor configured to;

receive a request to generate a candidate meal plan, wherein a seed list including one or more ingredients is designated by the cognitive system based on dish classifications, categories of ingredients, and cuisine classifications, wherein the one or more ingredients included in the seed list are incorporated into one or more recipes contained in the candidate meal plan, wherein the number of needed recipes to be included in the candidate meal plan is pre-determined;

import one or more recipes incorporating the one or more identified ingredients;

generate, by the cognitive system, one or more parent meal plans, each parent meal plan containing the pre-determined number of recipes, wherein the parent meal plan recipes are randomly selected from the identified recipes incorporating the one or more imported ingredients;

generate one or more child meal plans, wherein each child meal plan is generated by randomly selecting different parent meal plan recipes and combining the selected parent meal plan recipes;

wherein one or more mutations are strategically inserted by the cognitive system into the one or more child meal plans;

determine a fitness score for the one or more child meal plans through the utilization of a genetic algorithm, wherein the genetic algorithm comprises a fitness function that considers cost of ingredients, shelf life of ingredients, flavor compatibility of ingredients, waste ingredients, and preparation time as factors in determining the fitness score;

wherein the cognitive system generates an additional recipe using all the waste ingredients, and the additional recipe is incorporated into the one or more parent meal plans; and

output the candidate meal plan having the chosen number of recipes, wherein the child meal plan having the highest fitness score is selected to be the candidate meal plan;

wherein generating one or more parent meal plans, generating one or more child meal plans, and determining the fitness score for the one or more child meal plans are repeated for a pre-determined and finite number of iterations, wherein the one or more child meal plans with the highest fitness scores are used as the next generation of one or more parent meal plans.