You'll never get people to leave it till the optimum time to make use of it as a strategy
That's exactly why I put "letting" in quotes. I'm not saying that people should do anything different [1], just setting out the likely longer-term consequences of people doing what they're doing as the distances go down and the per-system difficulties go up. It's not as if there's been any deliberate strategy to let an Alert or two go to Invasion this week, after all.
Once enough close-in Alerts show up at once, they won't all be cleared at once, and then at some point so many will be going to Invasion at once that clearing all of
them in their first week will become impossible no matter how hard people try, so some will inevitably go to two weeks. That won't be a "strategy", it'll just be running out of bullets. That gives them the long four-week recovery time rather than the short one-week, which is enough to double the time until they get reattacked, which then serves as a partial mitigation of the fact that there's too many of them to start with. If people continue to push beyond
that point then eventually - as you say, invasions get tougher as they go on - systems will be lost not through any intent but simply through no-one having time to save them, and recapturing them will effectively be triple cooldown (though more likely at that point, not done at all).
If it was a
strategy it might be possible to pick the timings of that sort of thing to consistently reduce Maelstrom alert counts ... more likely it'll lead to boom/bust cycles where a Maelstrom - at least, the sparser ones - gets a temporary shell of long-Recovery systems it can't easily get past, so is restricted to uninhabited systems (or no Alerts at all) for a few weeks until that goes past ... and then it'll stick 5 high-difficulty inhabited Alerts down at once the next week and the cycle begins again with it being impossible to stop at least some of those going to multi-week Invasions no matter how much people try.
(As always, whether we actually get to that point before Update 17 or 18 changes things up, who knows)
Reducing Alerts seems to fluctuate, 25 a few weeks ago, now back up to nearly 40 again, so that clearly needs more work before any further tactics based on it can be formulated.
That's again mainly a consequence of variable cooldown times.
Knocking out a Maelstrom's entire outer perimeter - especially at the 20-30 LY ranges - in a couple of weeks appears to be feasible. That then puts all its 15-20 LY Controls facing a bunch of systems which are mostly on a four-week cooldown and no way to reach past, so Alerts temporarily and locally fall. Once those cooldowns are over, though, it's then got 15 targets (3-week cycle) perfectly readily and gets back up to speed. The tactic can't be repeated very often in general, because doing it again requires knocking out the 15-20 LY shell with similar speed.
Leigong and Thor are partial exceptions as their low density makes it hard for them to reliably get a 15-system perimeter in normal circumstances, and a few targeted control recaptures can push that down ... Indra, conversely, is going to get 5 Alerts a week pretty much whatever happens.
[1] I'm 22,000 LY away and have no relevant squadron affiliations. I'd have no ability to say that even if I wanted to.
