Question 332694
Stratified sampling
---
You randomly sample from "each" strata (ie 100 flight).  Each flight has 200 passenger out of a total of 100*200=20,000 so each strata makes up 200/20,000 =1/100 of the total.  Therefore 
sample from strata ith = total sample size *(prop strata is of total) = 800*(1/100) = 8 randomly selected from each strata
Keep in mind that if the strats had been different size, the proportions would be different and thus your sample make up from each strata would be different by this proportion
========
Cluster sampling
---
first randomly select a set of clusters then sample all the elements in the selected samples.  

So in your case, since each cluster is made up of 200 passengers, you need 4 random clusters to make your 800 sample.
========
Systematic sample
-----
there are 200*100=20,000 passengers, so you need to select every 20,000/800 =25th passenger. 

You need to begin at a random point, so you would select a random number between 1 to 25,
Then proceed sequentially from flight to flight (assume that each passenger in a flight can be identified as passenger 1 to 200), then begin by selecting flight #1, start with the random number you selected between 1 to 25 (say it was 8) select the 8th passenger in flight #1, and sequentially select the next one as passenger 33 (8th + 25) and then passenger 58th (33rd +25) move from flight to flight finding the passengers that match the 8+25*n multiple, from the original random number + 25*n.

===
each method has its applications, so some methods are more applicable than others.
Stratified sampling should be used when you suspect that the strata are hetergeneous (ie the "between" strata variation is very different but "within" strata variation is similar).
Cluster sampling should be used when you suspect that cluster are homogeneous 
(ie the "between" cluster variation is very similar but "within" cluster variation is very different).
Systematic sampling should be used when you have easily available listings, but is generally not the best type of sampling.

===
cost issues and administration complexities also play a role in making a choice of sampling.