Well at least in regards to distance from home, USA can use Britain as a base of operations as they did in reality. I agree it'd be tough but I still want the AI to give me a run for my money for once.
It took them til 1943 to win in North Africa, despite most german resources being directed elsewhere and sinking of axis supply convoys.
But I am gonna stop being a negative nancy and try to problem-solve instead, what we want is specialised AI behaviour for beating a player that has control over most/all of Europe. These tactics needs to be even more elaborate than what is needed to attack europe during historical 1943-1945 circumstances, since the player is probably far better prepared than the Germans were. Not sure to which degree this should be hard-coded because I have no idea how AI-design works. Alot of these have been written already, but here are my suggestions of special behaviors for this very occasion that could work:
1. Attack through a neutral country (sorry Portugal)
The advantage is obvious, how can those nasty krauts defend something they don't occupy, unseasoned portuguese divisions with 1936 rifles can be swept aside much easier than panzer battalions with veterans soldiers, allowing a foothold to be established while the player curses her timidity and respect for national sovereignty . It would also be helpful if the AIlies could make ultimatums to force neutral nations to enter the war on their side of face the consequences.
2. Attack through a relatively weak axis member/occupied territory
The soft underbelly of europe! (I realise this didn't actually work, but hey, the AI-lies should simulate the actual allies). Plus attacking the balkans or italy or spain (if they joined axis, otherwise see point 1
3. MASSIVE strength buildup
As someone mentioned (OP I think?) the allies did dniepr raid stuff but rarely did massive invasions, the AI-lies need to actually prepare a large force, and it has to be LARGE, not medium. If it's medium size force they end up loosing 10 divisions instead of 3. Their invasion needs to be kind of an all-in for it to have a chance for working. Also ridiculous amounts of air support needs to be employed.
4. Different targets from usual
High value should be placed on seizing ports and positions that would allow them to defend their bridgehead.
This might seems obvious but the AI needs to value getting a solid defensive position and getting access to ports much higher than they would under normal combat operations, and they need to be very ruthless when it comes to securing these places ("A kill zone you say? Well Ultra intercepts suggest they will run out of bullets before we run out of troops, so charge!")
5. Disruption
The AI-lies should try to combine their invasion efforts with coups, partisans and paratroopers to confuse the player. Both before, during and after initial landing. Beyond the obvious value of not having to assault the beaches and whatnot, this could help in other ways. For example, a coup in Romania will distract troops that were headed to defend Greece. A coup in Spain could convince the player the AI will be landing there when they were actually headed for Norway.
6. Fake plans
Ok so I read somewhere that if you have the right techs and/or luck you can get a hold of enemy war plans and vice versa. So there should be a focus or tech or decision or event or whatever that makes it so you can plant false plans, thus confusing the Player. They should also actually perform decoy invasions so that the player has to guess where the main strike is coming.
7. Mulberry Harbour(s)
Without the Mulberry Harbours the invasion of France could very well have failed, or atleast stalled. The Krauts, being ever so sneaky, destroyed harbours as they retreated (I can only assume a clever German player will do the same), forcing the Allies the rely on giant harbour-ships called mullberry harbours. The 2 harbours erected on the beaches of normandy would continue to be the main allied port in the west for 6 months, and it was used for 10 months in total. According to wikipedia some 2.5 million men and 4 million tonnes of supplies passed through it during that time. No harbour = no western front, no western front = Stalin whines even more about having to do all the work.
8. Higher degrees for mobilization
So historically the allies mobilized ALOT of their resources, but they never really went as close to total war as the Germans or Soviets. This is especially true for Americans since they didn't need all of their massive industry or manpower pool to do their part. For example (wikipedia once more) in 1944 the US barely outproduced the Germans in terms of AFVs, with 20357 vs 18956, and the German production included some 5000 panthers or tigers that were very expensive to produce compared to most american tanks.
However, with no other challenger to Germany in sight, an AI-merican dead-set on invading Europe needs to go all in, so that their massive industrial capacity and manpower reserve can compensate for the other unfavorable conditions. Perhaps a special focus only available after the USSR has been defeated?
The brits could get a similiar focus that allows them to draw on their colonies and the commonwealth (indian manpower?)
So if the AI-lies can do all these things, they might be able to put up a fight, or atleast go down in a blaze of glory.