Implement simple nectar production and consumption as a somewhat mechanistic foraging driver

@mlange @hanst This implements a simplistic foraging driver mechanism by limited the supply of nectar as real flower supposedly do. Note that I switched to point-like flower patches for now to avoid any question about shape until this becomes more interesting via actual GIS data.

@hanst Concerning our discussion on how to displace pollinators: I remembered reading something much simpler and much more in line with the IBM approach: Pollinators are supposedly very good at never visiting the same blossom twice as this would waste energy. Hence I just introduced a finite-capacity memory which flower patches were visited and avoid those until other ones are found. This seems both simpler and much more reasonable as pollinators cannot guess how much nectar is available until they actually try to get some. I will probably need to add some memory decay nevertheless, possibly coupled with sleeping?