For the situation where we have an automated algorithm, we run through a set of blends that you are willing to build, and we calculate the "cost" of over and or under application of using each blend.  If you want to make sure that you always apply the minimum need of either product, you set a high cost for underapplication.   If you want to cap the application to never exceed on either product, it is likely not possible if you have any zero application rate areas in the field for either product.

And... when you get done... you have to look at the cost for overapplication vs. the cost of separate application trips.