I’m en route to the March Meeting in Portland, which involves a three-hour layover in Chicago, between two flights on Southwest, my preferred airline. I’m always impressed by how much more efficient Southwest seems that the other major airlines.
One weird manifestation of that efficiency is the flight plans that Southwest uses. Where most flights on other airlines seem to go back and forth between two cities over and over, Southwest’s routes tend to roam all over the country. This morning’s flight from Albany to Chicago continued on to San Antonio, TX, Phoenix, and San Jose. Another recent trip involved a flight from Albany to Baltimore, Kansas City, San Antonio, Las Vegas, and San Jose.
This always makes me wonder about the computation behind these flight plans. There’s got to be some logic to it, that allows them to get all the planes they need to the places they need them, presumably as cheaply as they can manage. The result is most likely derived empirically from crunching some numbers on existing flight networks, then optimizing the results, but it’s kind of amusing to imagine that the folks at Southwest are sitting on a solution to the Traveling Salesman Problem, and using it to gain an advantage over their competitors…
(It’s also a good idea for a throw-away element of a science fiction story about Singularity-type stuff, but I’m not the guy to write that…)