Last game I played co-op with a friend of mine, I had almost all my sectors on energy focus due to the cost of keeping the fleet out of dry-dock for wars... And I ended up at like +80 food due to sectors. In fact I started building over most of my core world food tiles because it was just so stupid how much food I had.
Baffling to me is... when your empire is at +80 food, why the frick are the sector governors building orbital food modules onto the space ports? You only do that when you have a food shortage -- not when you have a massive surplus. Yeesh.
Right now I am playing a game with more core due to government and ascension perks and I am at 12 core systems, and about 14 worlds. No sectors yet. I am loving it.
Eventually when some worlds are built out, I will hand them off, but not before then.
Erm, what counts as a massive surplus depends on the number of pops you have, a good surplus is 0.4*pops, and a massive one would be >0.6*pops.
That aside the main issue appears to me to be the AI just being bad at deciding what it wants. I wonder, could the AI be expecting the growing pop to produce output, resulting in expected output code giving bad inputs that perpetuate a logic loop?
General rules the AI needs to follow IMO.
1. It probably needs a minimum excess energy and food and minerals (as a % of it's maximum maintenance capped at say 300 a month for EC and 500 for Minerals), and a desired Unity to worlds ratio.
2. It needs to build on tiles over redeveloping. It should only redevelop if it goes negetive on the 3 basics and there's no free pops.Or if all tiles are worked and have a pop on them.
3. Placing a building should set an AI flag on the affected tile. This prevent redevelopment for 10 years unless the AI is negetive and there are no free pops, or if condition 4 is met.
4. If a planet's base pre modifiers income of Minerals or Nexi exceeds 10 it should redevelop and existing mine or power plant into a processing plant or nexi.
5. If it needs to redevelop due to negatives it should redevelop resources in order of whichever is most over, (or nearest to if all below), the desired ratio and never redevelop any 3 basics if doing so would leave it negetive.
6. Whenever placing a desired building it should prioritise any tile with the maximum bonus that is free and has a pop on it.
7. It should never place building before a pop is grown and free. In effect only place buildings on tiles with pops.
8. Always run a pop reassignment check before redeveloping.
9. Never place space based stations if doing so would take it below ideal ratio.
10. Allow for different ideal ratios depending on fi standard or advanced start and number of years into game, (so standard starts can have very tight margins early on).
11. If all are above negetive, prioritise whichever is furthest from ideal ratio.
I can't think of anymore, but a basic setup like that, whilst not simple to code would do a lot to get the AI to mostly respect tile resources and not redevelop much.
