So now you have an idea of the versatility of an AESA system and why it is the default 'want' of every military.
With a large number of T/R modules, the partitioning will far exceed 9. And if the requirement is to track 64 targets at once, then how many such partitions do you expect an AESA to have?
Your question is best answered by understanding how the classical radar systems work in dealing with multiple targets.
As the antenna mechanically sweeps in volume search mode, if it detect any return, the system will remember the approximate locations of those returns. On the next sweep back, keep in mind that the antenna is sweeping back and forth, the system will momentarily change freqs at those locations. Higher freqs equal to improved target resolutions. Then as the antenna move away from those positions, the system switches back to volume search freqs again. The process is so fast that it give the impression that the system is doing several things at once. But it is a false or pseudo-multitasking capability.
With subarray partitioning in an AESA system, we have a true multi-tasker available. Keep in mind that there is a difference between antenna and array, the antenna contain the array. Say we divide the main array into two subarrays. We can have one subarray continue to perform nothing but volume search in the entire area in front of us and the other subarray does nothing but extremely fast switch its main beam from one target to the next. Because the two arrays are operating at different freqs, they will not interfere with each other. Neither subarray is multi-tasking, not even the array that is tracking multiple targets since all it does is move its main beam from one target to the next.
But what if the many targets are widely spread out? We can assign one subarray to one sector and the other subarray to the other sector, each performing that pseudo-multitasking of switching between volume search and target tracking. The advantage with a phase array system, PESA or AESA, is that because of the extremely fast beam movement, this pseudo-multitasking capability is much faster than with the classical mechanically motivated antenna. The downside with subarray partitioning is that because array size is proportionate to power output, the more subarrays we create, the less power output per array. The sum of all individual subarrays equal to original power output when the antenna was one array, of course.
With a PESA system, we remove the motors that move the antenna and move the single main beam, extremely fast, of course. With an AESA system, not only did we removed the motors, with subarray partitioning, we have multiple main beams to do as we please. The Russians installed motors with their ESA systems. That is not a bad idea. It give the system additional coverage thru mechanical mean. But with any mechanical system, there is an increased risk of lock-ups, stress, etc...etc...Take your chances.
There is no definitive figure to your question as the hypothetical situation of tracking 64 targets depends on the quality of the array choreography program, which cannot place any demands that exceed the quality of the hardware, which contains the T/R elements and how they were manufactured, the packaging of the related computers, the power supply, the shielding, etc...etc...
There are many strategies in writing array choreography programs. Each has advantages and disadvantages. We cannot demand the pilot to pick them out. He will be busy fighting for his life and for the greatest need of the conflict -- win the war. For example...Say we demand the pilot to create two subarrays, one for volume search and one to track 10 targets. But the nature of air combat is that nothing is static, not even the slow moving weather balloons and winds can carry them at a couple hundreds km/h, so what if 5 targets move beyond the scan limit of the tracking subarray? Do we program the system to automatically switch the two subarrays into pseudo-multitasking to keep track of all 10 targets? Or do tell the pilot to throw a switch or two to make it so? If you can imagine the arguments, pro and con, automatic or manual,
JUST FOR YOURSELF, now see how difficult it can be when many people have their own opinions just on how to deal with 10 targets with two subarrays while the system itself is capable of being far more than two subarrays.
Subarray partitioning and array choreography softwares are cans of worms, from technical issues to combat tactics, that have created a lot of animosities between people across many disciplines. It is wise to have a critical mind when coming across claims of accomplishments or capabilities from any manufacturor, even from US, or especially from the Russians and the Chinese.
:coffee_spray: