Well, given that the majority (more than 80%) of land is used for growing grass, wheat, soybean and corn for cattle there's no fear for vegetables.
Although soy grown for human consumption is not popular now, it's the most efficient legume there is. More protein than a steak.
This went a little bit off topic, although my point was that removing subsidies on food makes things tricky. Given that in some parts of US burger price is halved if not 20% of what it should be due to subsidies on water and corn and soybean etc.
Meat industry would collapse due to 1. Daily foods like steak, bacon, eggs, milk would get extremely expensive leaving the population to lousy unsatisfying vegetables.
Might solve the obesity epidemic :) Might also cause riots because people are not used to eating yucky foods like vegetabables.
It is not decision tree or any machine learning algorithm.
It is a local search algorithm that probably has simulated annealing or tabu search as a metaheuristic.
Although, they probably segment the deliveries to some common starting points and the problem size is reduced significantly - maybe to around thousand orders per starting point.
Research can easily optimize thousands of deliveries very effectively.