This has been going on for over 2 1/2 years now. At that time we had this item submitted:<p><a href="http://news.ycombinator.com/item?id=419360" rel="nofollow">http://news.ycombinator.com/item?id=419360</a><p>It looks like it's the same physicist, and the same algorithm. Further more, HN had pretty much exactly the same discussion.<p>Plus ça change, plus c'est la même chose.<p>There's another submission from over a month ago here:<p><a href="http://news.ycombinator.com/item?id=2790023" rel="nofollow">http://news.ycombinator.com/item?id=2790023</a><p>In that it's described how ...<p><pre><code> American Airlines undertook a two-year study to try and
speed up boarding. The result: The airline has recently
rolled out a new strategy—randomized boarding.
</code></pre>
I haven't seen any news of how that panned out.<p>This submission from 1300 days ago - <a href="http://news.ycombinator.com/item?id=111416" rel="nofollow">http://news.ycombinator.com/item?id=111416</a> - is a paper from Arxiv, suggesting that boarding times can be cut by a factor of 4. Guess who it's by - yup, our favorite physicist again. So he's been at this for 3.5 years. There are just 5 comments on that submission.<p>This latest paper is here: <a href="http://arxiv.org/abs/1108.5211" rel="nofollow">http://arxiv.org/abs/1108.5211</a><p>That was linked to from this submission: <a href="http://news.ycombinator.com/item?id=2943615" rel="nofollow">http://news.ycombinator.com/item?id=2943615</a><p>It was also referenced in the article pointed to in this submission: <a href="http://news.ycombinator.com/item?id=2943003" rel="nofollow">http://news.ycombinator.com/item?id=2943003</a><p>All in all, a popular topic that's been going for 3.5 years from this one physicist at least.<p>Despite his perseverance, it hasn't been adopted on any of the flights I've been on.<p>========<p>So here's a list of some of the previous HN items on this topic:<p><a href="http://news.ycombinator.com/item?id=111416" rel="nofollow">http://news.ycombinator.com/item?id=111416</a><p><a href="http://news.ycombinator.com/item?id=419360" rel="nofollow">http://news.ycombinator.com/item?id=419360</a><p><a href="http://news.ycombinator.com/item?id=924855" rel="nofollow">http://news.ycombinator.com/item?id=924855</a><p><a href="http://news.ycombinator.com/item?id=2790023" rel="nofollow">http://news.ycombinator.com/item?id=2790023</a><p><a href="http://news.ycombinator.com/item?id=2943003" rel="nofollow">http://news.ycombinator.com/item?id=2943003</a><p><a href="http://news.ycombinator.com/item?id=2943615" rel="nofollow">http://news.ycombinator.com/item?id=2943615</a>
Looking at the paper, it seems the passengers were the same and boarded five times in a row, with the last two methods proving the fastest. The major experimental flaw seems to be that the passengers themselves might learn to become more efficient after concentrating on a normally rare task, and repeating it several times in a row. Also, since some of them were paid extras, they might just be getting impatient by the end, and rushing through at what would normally be an uncomfortable pace. The population sampling is also probably not representative.<p>I would have found the experiment more convincing if it had been used to validate the basic assumptions of the theoretical model instead (e.g. the statistical distribution of the baggage loading and seating times).
I remember seeing this before.<p>I think it's really cool, though in the article I read before one major block to implementing this is that you'd be splitting up group boarding (of even 2 people travelling together).<p>I feel like that might be a tough message to try to explain to everyone at the airport, since in general people are worried about everyone in their party making it on the plane safely and with all their stuff. Gate agents have enough worried customers as it is.
He should totally patent his boarding method. Who cares that it could save the airlines billions? He could rake in so much dough by licensing the method or suing airlines who use his method without a license! And if the patent is vague enough, he could probably collect on all the other inferior boarding methods too!<p>In all seriousness, the boarding problem only got worse once airlines started charging for bags as people starting carrying on more and more. I read somewhere that Southwest actually saves more money by offering free checked bags and saving on boarding time than they would make had they charged for free bags.
Boarding by blocks starting at the front is ridiculous.<p>In fact, I'd be hard pressed to think of a worse way to board a plane. And yet somehow every time I fly that's how it happens. Maybe it's just that my company chooses horrible airlines.<p>The article mentions that assorted methods of boarding were tried, though it only goes into detail about "the Steffen method". I wonder what the difference between blocks-from-the-front and the obvious improvement of blocks-from-the-back is.
Airlines have been doing their own research into this, apparently. American Airlines is switching to randomized boarding (one of Steffen's proposed solutions) and United is partially switching to window-middle-aisle. <a href="http://online.wsj.com/article/SB10001424053111904233404576457930970524522.html" rel="nofollow">http://online.wsj.com/article/SB1000142405311190423340457645...</a>
My solution: Offer free drinks if everybody can board in <X minutes. Social pressures will do the rest (I occasionally see people help load heavy bags overhead, but I imagine this would pick up).<p>Not entirely joking!
I don't think this'll work. Airlines who have the most to gain from a short boarding time (budget airlines) have another approach, where they do not assign seats. Individual passengers are motivated to board quickly because they don't want to be sat next to a big fat person, or they want to stay with their group.<p>Ryanair has an average turn around time (time between when the plane lands to when it takes off again) of 25 minutes. I've been passed through the gate and waiting at the door before the plane I've to travel on has landed.
While we're on the subject of statistics, airplanes, and mathematically-but-not-socially-correct ways to do things, I really wish they'd have seats on the plane spaced according to a height distribution of passengers.
Savings from sophisticated method to board plane faster: 110 million dolars<p>Getting passengers to board when (and only when) it's actually their turn: priceless<p>(and, I suspect, far trickier)
I am interested to know where they got the savings estimate. $110M seems like a bit much for cutting the boarding time by 10 to 20 minutes. I was under the impression that most airline delays were caused by weather and traffic, not boarding times.
Imagine how much boarding time could be slashed if robots handled the baggage. The line stops moving when people are trying to store their bags, even if stepping aside for a moment would significantly reduce delays for others.
Boarding statused passengers first aside, it seems like this could be easily communicable with announcements of "boarding even rows", "boarding odd rows".<p>I skimmed the paper and didn't see any mention of how to get passengers to obey gate agent instructions, though, which would be a prerequisite of implementing an effective boarding method. Perhaps this should be re-tested by airlines randomly selecting sold out flights with identical planes to try these methods.
This looks good in theory, but if you have a party of 2 or more people, who are presumably seated together, then it is practically impossible to get them to board at different times as the algorithm would want to. I don't know for sure, but based on anecdotal observation I would guess that at least 50% of persons onboard travel with a co-passenger. I think that will throw the algorithm off quite significantly.
The article makes the assumption (or at least fails to demonstrate) that the bottleneck is <i>boarding</i>, as opposed to deplaning or cleaning. Other processes which may be done in parallel could just as easily be a bottleneck: food, sanitation, gas, luggage, and such.<p>And with super-fast turn-around times, airlines might have less time for inspections or maintenance, as we've seen with Southwest's recent problems. (Southwest reportedly has the fastest turn-around time in the industry, in part due to their standardized fleet.)
The delays I experienced is usually due to one or two passengers trying to get their last cigarettes before boarding.<p>Furthermore, there is no point getting too geeky about complicated boarding sequence if passengers are going to get unhappy over it.<p>Better to do a Steve Jobs and keep things simple.
I flew a few weeks ago and dialogued (argued? - tried to be good natured about it!) with the boarding staff at 3 different gates about boarding processes. I suggested they try windows first, then middle seats, then aisle seats. 2 out of the 3 argued back that the way they were doing it had been 'proven' by some study some years back by... either a Finnish airline, or some studies in Arizona - I honestly can't remember which they said (but they'd both said the same place). Very odd, because it's <i>demonstrably</i> pretty damn slow, and <i>often</i> the slowness is quite visible - people with window seats having to stop and climb over someone in the aisle and middle seats, causing a backup. Agreed, it's not the <i>only</i> cause of backups, but in my recent 6 flights, 5 of the boarding processes were rather significantly slowed by multiple window/aisle snafus.<p>As much as someone wants to say "we've studied this already, and this is the best way to do it!", you'd have a hard time convincing me that any major airline knows how to make good decisions about anything.
Menkes van den Briel did some work on airplane boarding. He has really nice explanations and videos here: <a href="http://leeds-faculty.colorado.edu/vandenbr/projects/boarding/boarding.htm" rel="nofollow">http://leeds-faculty.colorado.edu/vandenbr/projects/boarding...</a>
What a shock to read that despite this knowledge, the airline industry continues to do things the exact same way, wasting huge amounts of time and money and fuel in the process.<p>This really isn't a smart industry, in many respects.
this method seems to assume that people will line up in perfect order so that no one blocks anyone else, but that simply isn't the case. the alternating rows that were called would stagger in random order and blocking would occur as normal.
probably enforcing less carryon would speed the process up the most, but they'd have to make the checked in baggage pickup quick and easy to give people more incentives to check in.. maybe even offer a service that picks up rollies right at the gate and makes the available after getting off the plane..
It always puzzled me that a simpler algorithm has never been considered.<p>Passengers with heavier hand luggage or more than one piece (say hand-luggage and laptop) should be given back-seats. Anyone without hand luggage should be placed at the front seats.
Classic example of "cracking a nut with a sledge hammer". You don't need Monte Carlo to understand that boarding first rows is less efficient, or to come up with a much more efficient procedure. (And this is coming from a Monte Carlo lover).<p>What you do need is biz savvyness to understand that rows are ordered by passenger value (first and biz class first, followed by premium, platinum, frequent flyers, etc). These passengers pay a high premium to board and deplane first. Airlines are not going to lose these valuable passengers for a gain whose magnitude is uncertain at best.