Good points. And if you are going to model the behavior then model it. I am sure there are statistics on when families travel together or if a passenger is alone, on people running late, on failure to follow instructions and such. Go ahead and put it on the model and then you have a model where you can test the process and its sensitivity to these other variables that could screw it up.
I feel like I have said this before. And so did Wired.
I do like the genetic algorithm method he used to develop the boarding process and then test in the simulation. It seems like it wouldn’t be difficult to add the other factors.
