I've worked and designed a few FADEC systems so can add this RE how it computes mixture ...
first off. it's not like you setting a manual mixture every now and then....the electronic Engine controller , often genericlly called an EEC in airplane world, constantly adjust it based on tables, often called tunes in car world. these basic tables/charts have sensor inputs as ambient air temp (typically at/near air filter), ad MAP pres. more modern EECs also have barometric/altitude sensors, engine temp sensor , tachometer, and oxygen sensors in exhaust which can fine tune the tune.
the eec is not a AI device and thinks fr itself. it simply references all inputs, and the 'tune' table where all such inputs meet in the table, and commands the fuel injectors to open for a period of time (called pulse) to achieve that mixture ..... essentially, the table knows what millisecond pulse will achieve that exact mixture or air/fuel ratio based on the table pulse pre determined by the guy that configured the table.
a good tune is important for all operating conditions....if the table is perfect at low RPM, or for low inlet air temps, but off at higher rpm or MAP pres, the mixture will be off and could destroy engine. modern fadecs can also adjust spark plug timing if included in the tune table, which also can affect mixture, but that's another conversation.
knowing this now, and more directly answering your question, what really determines end result mixture ratio is how the tune was designed by the engineer that made it. for some applications it may be made to ROP. or LOP, or combo of each. for our c72 v8 tune we used LOP at cruise and load settings per general motors specs (not as lean as car tables are designed to run hotter cylinders for better emissions), and ROP on takeoff loads to help cool cylinders.
our designed used 2 EECs so we had to make our own tunes and what's called a calibration file (its like a computer operating system like Windows for your laptop).... but basic mixture ratios the same. I think you can see the Air/fuel ratio display on our YouTube video, and will see the AFR change based on RPM .... think of perfect mixture being 14.7 :1 (14.7 parts of air to 1 part fuel which should result in most efficient combustion)....if display is less its lean, above its rich.
we. as all others do, used algabratic formulas for starting tables which is close enough to get engine running enough to fine tune the table with just the basic air temp and MAP inputs, then add the other inputs and where we go more lean or rich. most is done on static dyno then more fine tuning in flight where dynamics as airspeed adds inlet pres, or prop loading changes effect the end result so the table again is more dialed in. a good tune will maintain a constant AFR as rpm, loads, temps change and a big part of flight testing. we enriched the mixture more with warmer air temps when above 2900 RPM to reduce cylinder temps and potential detonation burning takeoffs and go around where slower speeds and air inlet angles reduce cooling efficiency....our table has ability to make such complex changes as most modern EFI systems. older or more basic eec may only have the min required imputed as temp and MA.
Modern EECs may also have an anticipatory circuit that figured if you quickly go from low power to mac, as if shoving throttle in quiclias in go around it anticipates you are jumping to high rpm/load table and jumps to it before sensors react.
so to you question, fadec tune determine mixture and a function of not only how many inputs a table has (usually limited by EEC processor), but the exact use application and real time trial and error. I can't make blanket statement that all tunes run same normally leaner ratio models, but typically tables and calibrations are usually configured to run slightly lean of peak based on objectives as emissions, fuel savings, cylinder temps and detnation, loading, etc.
hope this helps
