If we're looking at this being a long term problem, we should be looking at long term solutions.
Long term solutions should be considered on a risk management basis and as soon as you move down this path, you have to consider what is an acceptable failure rate.
If 1 in 10 people are not capable of self isolating correctly and 1 in 10 of people arriving from overseas are carrying the virus, you could combine this with a potential 1 in 5 risk of their isolation transgressions spreading it and you end up with a 1 in 500 chance that an overseas arrival, self isolating, will cause a cluster. Not very good odds, all things considered.
Introduce good policing and follow up checks, it seems reasonable that you could drop that isolation failure to 1 in 50, possibly better. Good scrutiny, testing and discussion with the incoming traveller could easily drop that risk that they're infected to 1 in 100. That brings you back to 1 in 25,000 that they'll spread it.
Generally 1 in 100,000 is considered acceptable in industry, but that's for fatality, not spreading a virus. Last I heard MVA's were at 1 in 14,000 but we still drive. Hence the question; what's acceptable?
Considering it this way may seem pedantic but we're fighting a mathematical situation.
Unfortunately, the worst people to be handling this are politicians. They can do the maths better than most but their sums include a lot of factors that straight public health benefit equations do not!