I can understand why you might think those variables are worth a try, though I’m going to try to encourage you to re-think that.
First, let’s tackle the dose doubling. From our earlier mean plasma drug levels example, we know that the mean of a 10mg dose is ~120 ng/ml. Zolpidem (Ambien) follows what we call linear dosing kinetics, at least up to 20mg, which is a fancy way of saying if you double or triple the dose you give to someone you roughly double or triple the maximum concentration. If you were to take 15mg, this means you’d be increasing the mean by 1.5 times, so 180 ng/ml, and at 20mg, 2 times, so 240 ng/ml. This doesn’t actually buy you all that much extra time asleep under the drug’s influence, because the mean half-life is still 2.5 hours. If we do 20mg, you peak 1.5 hours in at 240 ng/ml, 2.5 hours later (4 hours post dose) you’re down to 120 ng/ml (same as if you had taken a single 10mg tablet), and go from there, meaning for taking 2 tablets, you roughly get an “extra” 2.5 hours of sleep over the 6-7 hours you got from a single tablet (in somewhat idealistic conditions, to be sure). Seems not so worth it, if you ask me, to get less than half the benefit from doubling the dose (this phenomenon where progressive increases in drug doses produce smaller and smaller benefits is not unusual for many drugs), especially since side effect risk goes up (being almost unrouseable in an emergency situation on a plane being one possibility).
Alcohol, of course is known to produce several actions, depending on the BAC levels you reach, with effects being stronger as the concentration is going up versus coming down, meaning that alcohol at, say, a BAC of 0.08 with the curve still rising is going to have more profound effects on mental function than the BAC of 0.08 with the curve going down (this is known as the Mellanby Effect). Taking your Ambien, especially at the higher doses you are considering, will likely coincide with rising/peaking ethanol levels, depending on how fast you drink, and the effects are at the very least likely to be additive if not somewhat multiplicative on each other. Alcohol, at a BAC of 0.08 has a concentration of 17.4 millimolar (~5 millimolar is all that is necessary for some degree of impairment to occur), on it’s own. Potentiate, whether additively or multiplicatively, with Ambien, and you run the risk of much greater disinhibition (talking loudly, expansively, aggressively) while blocking memory formation as, at sufficient doses, Alcohol enhances tonic GABA-A firing (via supersensitive GABA-A receptors containing an alpha-4 and delta unit in parts of the brain including the dentate gyrus of the hippocampus), while zolpidem enhances phasic or burst GABA firing at at GABA-A receptors located synaptically with an interfacing alpha-1 and gamma-2 unit, strongly believed to be involved in memory formation, among other things. Alcohol also enhances various potassium channels (which ultimately slows neuron firing down), voltage-gated calcium channels (involved in neurotransmitter release, both excitatory and inhibitory), and inhibits NMDA function all in doses which are well below fatal (NMDA being involved in synaptic long-term potention believed to be necessary for at least some memory formation). Point being, you’re now messing with a hell of a lot of brain circuitry with just two drugs, with no way to know (until computational modeling gets way better) just exactly what the results will be. Oh, and the combo can stop your breathing altogether.
The flipside is there are plenty of sleep drugs which appear to have different safety profiles which may not put you at risk for some of the effects mentioned in this thread. Given that insomnia is often the most refractory effect to medications in disorders like depression and anxiety, you might actually benefit from something like low-dose doxepin (a tricyclic antidepressant in high doses, but which has relatively pure anti-histaminic properties argued to be without anti-cholinergic properties at doses of 3-6mg at bedtime, with a short enough half-life to minimize next morning impairments, to the point that an author of Essential Psychopharmacology suggests tolerance may never develop to it’s effects (since it, unlike benadryl, lacks the aforementioned anticholinergic side effects at sedative doses). And while the low dose form is currently brand only (Silenor), a simple work around is to have the doctor write for doxepin 10mg/ml concentrate, and just use a 1 ml oral syringe to dose 0.3 to 0.6 ml (diluted in a small amount of beverage). A 4 ounce bottle costs roughly $27.99 and would last you over 200 days at 5mg (0.5 ml), without insurance. Likewise, really low dose mirtazapine (1/4 to 1/2 of a 7.5mg tablet) seems to produce the same effect, albeit with a much longer half-life. Then there’s trazodone, which, by combining certain effects (alpha-1 antagonism and anti-histaminic properties) may produce adequte sedation to last one through the night (though caution should be used in individuals on multiple anti-hypertensive agents or whose blood pressure runs normally low), and may serve as augmentative therapy for not-fully controlled depression.
tl;dr - Don’t let the side effect talk scare you too much. If you don’t want to try Ambien, Lunesta, Sonata, or any of the benzodiazepines, there are plenty of other options your prescriber can try to alleviate insomnia symptoms, including non-drug therapies such as cognitive behavior therapies.
Tramadol wouldn’t be my first choice for restless legs syndrome. Preferred options are currently the dopamine agonists (pramipexole, ropinirole, carbidopa/levodopa) or alpha-2-delta ligands (gabapentin and Lyrica) for chronic RLS, with tramadol probably being safer than other opioids, particularly in acute instances. Moreover, there is at least one case report that tramadol-withdrawal actually precipitated a bout of RLS–other reports for other opioids have also been observed (see here for the case report, and here for an abstract on current preferred treatment options). if your prescriber already tried the preferred options, of course, than yes, tramadol at 100mg once a day is probably preferable to a stronger agent like oxycodone.
Limiting one’s mobility on a long flight puts one in particular danger for the development of dangerous blood clots, especially if the patient in question is completely out of it and not getting up and moving around every few hours, as would be expected under sedative-hypnotics, particularly at higher doses, and/or in combination with ethanol. As callous as it may sound, a night or two of less than optimal sleep due to airplane travel generally sucking these days may be worth their own risks versus the risk of a memory impaired dis-inhibitory reaction with ambien alone, in high doses, or mixed with alcohol. Compression stockings may help (but are no guarantee) on the blood clot front, but at least you’ve been informed of (some of) the risks with what you are proposing. Ultimately, you are your own advocate and you have to make the choice for yourself.