Just to catch up with this before it becomes necromancy;
I was thinking of a heavily repulsor- based order of battle partly for high speed manoeuvre in open field conditions, but also ecumenopolis or major city targets. I doubt a treaded or walker vehicle is going to be particularly good at dealing with freedom fighters on the three hundredth floor...hm, there's a cheesy movie to be made in there somewhere. No matter.
The population figure, there is definitely space infrastructure, there are skyhooks around Coruscant, there's Centrepoint, there are the yards at Kuat, the Star Forge over Lehon, there's Bespin- but all of the obvious examples (apart from the skyhooks, I'll admit to those)seem to be there for some industrial or economic purpose. They're not empty, but they're also not primarily habitats.
Why assume that all species are equally- or at all- fortunate? There are twenty million sentient species, and frankly most of them aren't that impressive; Ewoks? Gungans? Why can there not be only a few billion of each of them on average, some with more, some with even less? There are a lot of less developed worlds in the mid and outer rim, and a concentration of major worlds in the core. I think a lopsided distribution of population fits that.
Is it not possible that the member species of the Trade Federation have economic power out of proportion with their numbers? They were heavily involved in manufacture and banking, and used largely droid automation for the former.
On the Aethersprite, two interesting comments, the quad cannon are explicitly stated to be 1 kiloton per shot maximum (nothing about rate of fire though) and
Tarriff barriers and embargoes between the galaxy's sectors prevent direct competition[...] in years to come Kuat will gain a near monopoly on warship contracts, although rival Sienar will win most government business for starfighters. Thus the technically excellent Delta-7 will be extinct in ten years...
That suggests that the Aethersprite was killed by politics far more than by performance- but it still raises the question, why was the TIE fighter considered a worthy replacement? Were the political imperatives so strong that any piece of crap would do as long as it came from Sienar, or was it close enough that the politics simply tipped the balance?
If there's something other than firepower that the TIE/ln might be better than the Aethersprite at, I can't think of it. Actually, the Aethersprite would be a superb surface support platform- small, agile, tricky to hit, enough firepower to kill most surface vehicles- and TIE series starfighters, probably.
Back on topic, my own take; I started from the regimental level intending to work outwards in both directions, and ended up with a division containing four combat regiments plus support.
Divisional assets, workshop and foundry droid pool, field hospital, counterorbital artillery (four firing batteries of eight convergence beam artillery turbolasers on heavy repulsor chassis), two squadrons of fighter-bombers, TIE/sa if necessary, ideally Hunter, Starwing or, the only Rebel type I would consider prying from their cold, dead hands, the X-wing (originally offered for sale to the Imperial military anyway.)
Each regiment, heavy armour batallion, light armour batallion, mech infantry batallion, cavalry batallion, artillery batallion. Fill in command and control, logistics, staff elements- they ought to be there. The assumption is that the unit will be cut off and thrown back on it's own resources, so they take at least two months' spare parts and stores with them. Shouldn't happen, but plan for the worst.
The heavy armour batallion consists of four companies of eighteen (four platoons of four plus command, exec) gunships as theorised earlier.
The light armour batallion of four companies of twenty- six (four platoons of six plus command, exec) T-47 or upgraded version, rebel airspeeders, with armament changes. Strip the tow harpoon, replace with PLX or similar. Forward, in place of the twin medium blaster cannon, either upgrade to higher power, lower rate of fire laser, grouped light blaster cannon to spray light vehicles and infantry, or convergence beam ball turrets, remote controlled from the gunner's seat.
The mech infantry batallion aren't really; something like the existing armoured personnel carrier seen in Rogue Squadron III, although heavily updated and uparmoured- I like the look of the thing but it's performance seems to have been pathetic.
Each carrying a payload of droids rather than human or other live infantry- urban operations are just too damned dangerous. Four droideka if they can be made cheaply enough, four ASN-121 (Zam Wessell's droid) with something more like an actual weapon- fit them with an E-11 instead of that silly canister/worm thing, and eight foldaway B1 battle droids.
4 companies of 18 (4 to a squad plus command and exec).
Cavalry are one and two man repulsor vehicles, two companies each of 64 two man swoops, pilot and gunner with pintle mount infantry support weapon (each 4 platoons of 4 squads of 4 units), possibly with an assassin-droid dismount, two companies of speeder bikes, 4 platoons of 4 squads of 8 bikes, half equipped with a small cannon for use on terrain and light vehicles, half with an infantry weapon on an eyeball sight.
Regimental artillery support batallion would have what mobile shields are available attached to it, it's the closest an almost entirely mobile force has to a solid base.
Tactical missile fire can be done by the gunships, and missiles are likely to be more use in an urban environment given the tech level- they seem able to dodge round obstacles on their way to the target, meaning less collateral damage and less wasted ammo. Very heavy artillery are kept at divisional; what about a mix of mass driver tubes (smart shells of course) and navy standard light turbolasers? Aerospace defence if the mount is anything like up to the same standard, firepower sufficient to cope with most heavy things that get in the way. Two companies- each four platoons of four- of each.
