New research shows letting the slowest passengers board first leads to the most efficient boarding overall. Look for the followup study, “How Exasperating Sighs and Eye-Rolling Slow Plane Boarding.”
Trying to determine the best way to execute a chaotic activity like plane boarding is an optimization problem: Scientists choose parameters to try to minimize or maximize, like the time it takes for everyone to board the plane. These problems are very old, but computing technology has given scientists powerful new tools to “brute force” solutions, meaning they can run long lists of complex calculations that would have taken human mathematicians years or even lifetimes to do by hand.
Imagine trying to guess someone’s four-digit PIN by entering all possibilities, one at a time, into an ATM. A computer plugged into the ATM could spam all the combinations much faster and more directly. (This is one reason why the three-tries limit is so common.)
For their simulations, researchers in Norway, Israel, and Latvia used a four-dimensional model of spacetime. This approach, called a Lorentzian metric, uses the same principles that inform general relativity and lets researchers model whether passengers are delaying each other in line (“timelike” separation) or able to board simultaneously (“spacelike” separation). They posited that the fastest passengers are five times faster than the slowest and used a congestion parameter set at three different levels to gnarl the process in a realistic way.