Where does the AI go wrong with districts?

[Sostratus]

I think the core problem here is "tiles are not there just for districts" - a District Placement Plan will (be) interfere(d) with and by any other priority about Wonders and improvements. Actually, it may already be that way (eg idle builders).

This particular issue is always present in any system where the computer needs to place districts, including the ones we have now. This is also why I had suggested that a city could be partially pre-planned on settlement which by default means that
1) Any city settled before that already has its districts planned on the map
2) Any city evaluating a placement decision of {district wonder improvement} after this computation will take into account whatever this city has planned
The only way for such a plan to be interfered with is if

• sea levels rise
• a strategic resource pops
• another civ takes the tile or conquers the city
• a city is settled and that tile lands in its first ring

but that's why a good algorithm (really just a procedure to make the best of an arbitrary input, algos aren't anything magical) has ways to deal with that. in my example, if the tile is judged permanently invalid, it's booted from the queue. But I also left the current placement mechanism in there so that it can do business as usual if it comes back to it. The only real change to what's happening now is that if a city wants to place district X and district X has been planned, then place it where it was planned. Currently the same procedure occurs except it's evaluation of where to place it happens at the same time instead of earlier.

 

[AsH2]

This particular issue is always present in any system where the computer needs to place districts, including the ones we have now.
..

Well, as it's now it's not fail-safe.
What I suggested, in the part you left out, would not suffer in permanent placement issues of both districts and wonders.

Don't get me wrong, I do enjoy this new District concept, but it's so backward and it cripple humans planning though mostly AI - there need to be a better system where all numbers are in place when crunching.
A district should be generic at first and later earn a specified status, when it's making sense - when there's infrastructure for it (@Boris Gudenuf in Industrial Era?).
Until then adjacency should be about buildings in a district - buildings would be (re)movable somehow, in a project (when there's technology/civic for it).