Determining Optimal Fertilizer Mixes
[wtm_use_case_cats]
A farmer will provide a company sales representative with information about previous crop yields and his or her target yields and the company representatives will visit the farm to obtain soil samples, which are analyzed in the company labs. A report is generated, which indicates the soil requirements for nutrients, including nitrogen, phosphorus, potassium, boron, magnesium, sulfur, and zinc. Given these soil requirements, company experts determine an optimal fertilizer blend, using a linear programming model that includes constraints for the nutrient quantities required by the soil (for a particular crop) and an objective function that minimizes production costs.
Benefits for the company
Previously the company determined fertilizer blend recommendations by using a time-consuming manual procedure conducted by experts. The linear programming model enables the company to provide accurate, quick, low-cost (discounted) estimates to its customers, which has helped the company gain new customers and increase its market share
Feasability
Medium
Type of expertise/ AI domain
Linear Programming Model and Operations Research
Internal data required
Soil Type, Soil Constituent
One of the fertilizer manufacturer and leading producer and distributor of specialty fertilizers in the world, with revenues of almost US$ 2.0 billion in more than 80 countries. XYZ produces seven main specialty fertilizers and more than 300 fertilizer blends, depending on the needs of its customers. Farmers want the company to quickly recommend optimal fertilizer blends that will provide the appropriate quantity of ingredients for their particular crop at the lowest possible cost.