Hybrid seed selection and seed portfolio optimization by field

Abstract

Systems and methods are provided for generating a set of target seeds with optimal yield and risk performance. One example computer-implemented method includes generating, by a server, representative yield values for a group of seeds based on historical agricultural data, generating a dataset of risk values for the seeds, and generating a dataset of target seeds from the seeds for planting in one or more target fields based on: the dataset of risk values, the representative yield values, and properties for the target field(s). The method also includes generating, by the server, allocation instructions for the target seeds included in the dataset of target seeds, where the allocation instructions are indicative of, for each target seed, a planting quantity for the target seed and a planting location for the target seed within the target field(s).

What is claimed is: 
     
         1 . A computer-implemented method comprising:
 generating, by a server, representative yield values for one or more seeds based on historical agricultural data;   generating, by the server, a dataset of risk values for the one or more seeds, the dataset of risk values indicative of an amount of risk based on yield variability associated with the representative yield values for the one or more seeds based on the historical agricultural data;   generating, by the server, a dataset of target seeds, from the one or more seeds, for planting in one or more target fields based on: the dataset of risk values for the one or more seeds, the representative yield values for the one or more seeds, and one or more properties for the one or more target fields;   generating, by the server, allocation instructions for the target seeds included in the dataset of target seeds, the allocation instructions indicative of, for each target seed in the dataset of target seeds, a planting quantity for the target seed and a planting locations for the target seed within the one or more target fields; and   causing display of, on a display device, the dataset of target seeds including: the representative yield value of each of the target seeds, the amount of risk for each of the target seeds, and an indication of the one or more target fields.   
     
     
         2 . The computer-implemented method of  claim 1 , further comprising, prior to generating the representative yield values:
 determining a dataset representative of multiple candidate seeds, the dataset including probability of success values associated with each of the multiple candidate seeds, which describe a probability of a successful yield, and historical agricultural data associated with each of the multiple candidate seeds; and   selecting the one or more seeds as a subset of the multiple candidate seeds, the one or more seeds having probability of success values greater than a target probability of success filtering threshold.   
     
     
         3 . The computer-implemented method of  claim 2 , wherein the probability of success values indicate a probability that a yield value of a respective one of the candidate seeds exceeds an average yield value of other ones of the candidate seeds based on historical agricultural data. 
     
     
         4 . The computer-implemented method of  claim 1 , wherein historical agricultural data comprises annual yield output as bushels per acre and seed relative maturity. 
     
     
         5 . The computer-implemented method of  claim 1 , wherein the one or more properties for the one or more target fields comprise geo-location and size of each target field in the one or more target fields. 
     
     
         6 . The computer-implemented method of  claim 1 , wherein the allocation instructions are generated based on a sum of the representative yield values for the target seeds included in the dataset of target seeds and a calculated sum of risk values for the target seeds included in the dataset of target seeds that is below a configured total risk threshold. 
     
     
         7 . The computer-implemented method of  claim 1 , further comprising causing display of, on the display device, the allocation instructions for each target seed in the dataset of target seeds. 
     
     
         8 . The computer-implemented method of  claim 1 , wherein generating the dataset of risk values for the one or more seeds includes calculating a year-over-year variance risk of yield values for each seed of the one or more seeds as a variance of yield values over two or more years for a specific seed based on the historical agricultural data for the specific seed. 
     
     
         9 . The computer-implemented method of  claim 1 , wherein generating the dataset of risk values for the one or more seeds includes calculating a field-by-field variance risk of yield values for each seed of the one or more seeds as a variance of yield values from two or more fields for a specific seed for a specific year based on the historical agricultural data for the specific seed. 
     
     
         10 . The computer-implemented method of  claim 1 , wherein generating the dataset of target seeds includes:
 using an optimal frontier curve to generate a specific target threshold for the representative yield values for a range of the risk values included in the dataset of risk values; and   selecting the target seeds that have representative yield values that satisfy the specific target threshold for the range of risk values from the dataset of risk values.   
     
     
         11 . An agricultural intelligence computer system comprising:
 a processor and a non-transitory computer-readable storage medium, which includes instructions which, when executed using the processor, cause the processor to:
 generate representative yield values for one or more seeds based on historical agricultural data; 
 generate a dataset of risk values for the one or more seeds, the dataset of risk values indicative of an amount of risk based on yield variability associated with the representative yield values for of the one or more seeds based on the historical agricultural data; 
 generate a dataset of target seeds, from the one or more seeds, for planting in one or more target fields based on: the dataset of risk values for the one or more seeds, the representative yield values for the one or more seeds, and one or more properties for the one or more target fields; 
 generate allocation instructions for the target seeds included in the dataset of target seeds, the allocation instructions indicative of, for each target seed in the dataset of target seeds, a planting quantity for the target seed and a planting location for the target seed within the one or more target fields; and 
 cause display of, on a display device, the dataset of target seeds including: the representative yield value of each of the target seeds, the amount of risk for each of the target seeds, and an indication of the one or more target fields. 
   
     
     
         12 . The agricultural intelligence computer system of  claim 11 , wherein the instructions, when executed using the processor, further cause the processor to, prior to generating the representative yield values:
 determine a dataset of multiple candidate seeds, the dataset of multiple candidate seeds including probability of success values associated with each of the multiple candidate seeds, which describe a probability of a successful yield, and historical agricultural data associated with each of the multiple candidate seeds; and   select the one or more seeds as a subset of the multiple candidate seeds, based on the one or more seeds having probability of success values greater than a target probability of success filtering threshold.   
     
     
         13 . The agricultural intelligence computer system of  claim 12 , wherein the probability of success values indicate a probability that a yield value of a respective one of the candidate seeds exceeds an average yield value of other ones of the candidate seeds based on historical agricultural data. 
     
     
         14 . The agricultural intelligence computer system of  claim 11 , wherein historical agricultural data comprises annual yield output as bushels per acre and seed relative maturity. 
     
     
         15 . The agricultural intelligence computer system of  claim 11 , wherein the one or more properties for the one or more target fields comprise geo-location and size of each target field in the one or more target fields. 
     
     
         16 . The agricultural intelligence computer system of  claim 11 , wherein the allocation instructions are based on a sum of the representative yield values for the target seeds included in the dataset of target seeds and a calculated sum of risk values for the target seeds included in the dataset of target seeds that is below a configured total risk threshold. 
     
     
         17 . The agricultural intelligence computer system of  claim 11 , wherein the instructions, when executed using the processor, further cause the processor to cause display of, on the display device, the allocation instructions for each target seed in the dataset of target seeds. 
     
     
         18 . The agricultural intelligence computer system of  claim 11 , wherein the instructions, when executed using the processor, further cause the processor, in generating the dataset of risk values for the one or more seeds, to calculate a year-over-year variance risk of yield values for each seed of the one or more seeds as a variance of yield values over two or more years for a specific seed based on the historical agricultural data for the specific seed. 
     
     
         19 . The agricultural intelligence computer system of  claim 11 , wherein the instructions, when executed using the processor, further cause the processor, in generating the dataset of risk values for the one or more seeds, to calculate a field-by-field variance risk of yield values for each seed of the one or more seeds as a variance of yield values from two or more fields for a specific seed for a specific year based on the historical agricultural data for the specific seed. 
     
     
         20 . The agricultural intelligence computer system of  claim 11 , wherein the instructions, when executed using the processor, further cause the processor, in generating the dataset of target seeds, to:
 use an optimal frontier curve to generate the specific target threshold for the representative yield values for the range of risk values; and   select the target seeds that have representative yield values that meet the specific target threshold for the range of risk values from the dataset of risk values.