# The Best Way to Board a Plane

Most airlines board passengers the same way, first filling the seats in the back of the plane, and then moving to the front. After a recent experience boarding a plane in this manner, Fermilab physicist Jason Steffen wondered if there might be a better way. So, in the midst of studying gravitation and axion-like particles, Steffen took a short break to investigate an optimal boarding method for airline passengers.

“I remember waiting in line to scan my ticket inside the terminal, I believe it was at the Seattle airport,” Steffen told PhysOrg.com. “I remember being quite disappointed when I saw how long the second line was – the one at the entrance to the airplane – and how slowly it moved. . . . That's when I thought that there had to be a better way to get people onto the airplane than the one that was being employed. I didn't have the time to work on it right then, so I brooded over it for almost 18 months. Last year, I decided that I either needed to solve the problem or stop thinking about it.”

In his analysis, Steffen found that the worst method for boarding a plane is boarding from the front to the back, since passengers have to wait and step over each other to get to their seats. As he explains in a paper submitted to the Journal of Air Transport Management, conventional wisdom suggests that boarding in a manner opposite to the slowest method seems like it should be the fastest method. Quite unexpectedly, then, Steffen found that the common back-to-front boarding method is actually the second worst method possible, only slightly better than boarding front to back.

“I was certain that the worst way to load the airplane was from front to back, so I ran my simulation in that configuration first to set a baseline,” Steffen said. “I was also somewhat convinced that the optimal way would be from back to front or something like it. I half expected to find that back-to-front loading is several times faster than front-to-back. Had that been true, I was prepared to run the two simulations, see how much faster it was, be satisfied, and put it aside. When the results were almost identical, I first thought that there was a bug in my code. Once I was convinced that my code worked properly, I realized that the problem was more interesting than I had anticipated, and I got more serious about it.”

Using a combination of a Monte Carlo optimization algorithm and intuition, Steffen determined an optimal boarding method, which could make boarding go 4 to 10 times faster than the worst method, depending on the size of the plane. In the optimal method, passengers would board 10 at a time in every other row (since loading luggage requires about two aisles of space). This way, passengers could always be boarding luggage or sitting in their seats, rather than waiting in the aisle, as in the two previous methods.

However, Steffen also acknowledged that the optimal method might not be practical, since passengers who sit next to each other often travel together, and prefer to board together. He proposed a modified version, where passengers board in blocks of three consecutive seats on one side of the plane in every other row. In this strategy, there would be four boarding groups, with passengers in the same row on the same side boarding together. This method provided a decent middle ground, as it was twice as slow as the optimal method, but twice as fast as the conventional method. Although getting passengers to line up in their correct groups might sound challenging, Steffen noted that Southwest Airlines has been experimenting with having its passengers line up in numerical order – so the logistics wouldn’t be inconceivable.

Steffen also identified several other boarding strategies with results superior to the conventional method. Contrary to our tendency for order, even completely random boarding proved to be a good alternative. In fact, random boarding was nearly as fast as the modified optimal method. Plus, by its very nature, it has the advantage of not requiring airline attendants to organize boarding passengers in any way. And the random result also shows that, when passengers board out of order in the other strategies, the results will still be better than the conventional boarding method.

The main advantage of the alternative boarding methods is that they allow several passengers to load their overhead luggage simultaneously, which Steffen identified as the largest factor in determining boarding time. By spreading the passengers throughout the airplane instead of concentrating them together, more passengers could load their luggage at once. Steffen noted that, although he has recently heard of other boarding optimization studies, his analysis uses a unique method and is the first to generate this specific optimization strategy.

“I think that the biggest challenge to implementing one of these methods is cracking into the industry,” he said. “Right now, I have a model where the parameters need to be calibrated with data. But that would require an investment from an airline company or manufacturer. While I could be wrong, I doubt that when an airline company needs to study an issue like this one that their first thought is, "Let's go talk to a physicist" (followed by, "Look, here's one that's studied axions and extrasolar planets. He's our man.") The two fields just don't talk to each other enough to have that kind of understanding.”

Still, Steffen thinks that reducing the boarding time could benefit airlines in a number of ways, especially for short flights between nearby cities. In such cases, quicker boarding might allow an additional daily flight to be scheduled, or it could reduce the number of gates an airline requires, since each gate could be cleared more rapidly. With thousands of flights taking off around the globe every day, a few minutes could save a lot of people a lot of time.

More information: Steffen, Jason H. “Optimal boarding method for airline passengers.” ArXiv:0802.0733v1 [physics.soc-ph] 6 Feb 2008.

All rights reserved. This material may not be published, broadcast, rewritten or redistributed in whole or part without the express written permission of PhysOrg.com.

Feedback to editors

Feb 14, 2008
If speed of passenger loading is the singular goal, there are many other variables that one should consider. Like requiring all large baggage to be checked. Or using both the rear and the front doors. Or providing larger, easier to load overhead storage. Or interspersing "porters" through the passenger line to expedite baggage loading. Or create one large compartment near the entry door for carry-on baggage(some airlines have done this in first class). Or simply allow more room for people and baggage. Ignore that last one, as I'm afraid airline greed trumps all.
And the article failed to give results on the old favorite: windows board first, aisles board last.
I think there are more simulations to be run.

Feb 14, 2008
I'd like to meet this physicist ;-)

Feb 14, 2008
windows board first, aisles board last.

It's only true if they don't have any lugguge, IMO.

Feb 14, 2008
Mayday, I think the idea was to have the largest increase in efficency using the smallest number of changes. Airlines aren't going to pay extra people to be porters, and *no one* will want to check all their baggage, besides the fact that it would require more baggage handlers to be hired, which the airlines won't do. (I temporarily lose my baggage 1 out of every 3 flights. Maybe I'm unlucky, but I'm darned glad that I carry a pair of socks, underwear, and toothbrush/toothpaste (etc) with me in my carryon! Fortunately I've never lost it for good. Its always been sent to me within 3 days (often the next day).)

Just thinking about this, and without running simulations, I'd say that boarding every 2nd row window seats first (with exceptions made for people travelling in groups, which the airline could automatically note and account for) would be the optimal solution. I can't see that causing any headaches for either the airlines or the passengers.

And, if boarding efficiency is increased, they could potentially increase the efficiency of the entire airport. Assuming that it isn't a hub airport. In those airports the runways and approach/takeoff zones are the bottlenecks.

Feb 14, 2008
I think the proposed system sounds delightful. Fewer people per section means people will have room to load their luggage (if they have any) and seat themselves. Letting multiple sections board at the same time means more people will be able to benefit from this lack of crowding. It would take the average person less time to get seated and more people could be seated at the same time.

The system would not be very difficult to implement. The airline companies should take notice.

Feb 14, 2008
There are two problems here:

1. people who never fly
2. people who fly too much

Both issues can be addressed in the same manner:

1. If you are in the aisle, you are in the way and keeping EVERYBODY from getting there on time. (This is directed both groups.)

2. People who never fly: see #1.

3. People who never fly: if you don't know what to do, ask. Also, look up to find the current row number: many times the row number "jumps up" to a higher number in the "coach" area... If you didn't buy a FIRST CLASS ticket.

Feb 14, 2008
Get this news to the airlines, pronto.

I hate that.

Feb 14, 2008
If random boarding is better than the current system, try that first. It's just common sense, and of course it costs nothing extra.

Of course, the first classies will want to board first because they paid the extra money. So get them out of the way, then start optimizing.

Oh and of course handicapped passengers and those with small children. Let them through first too.

Then tell the rabble, "randomize yourselves or get thee directly to Guantanamo." Maybe some of them will just run away. Wait five minutes, then take off, use it or lose it and all that.

The Bush administration will side with the airlines, and then courageous Democrats in Congress will first declare their opposition, and then compromise. Completely.

All aboard!

Feb 15, 2008
I work as ground crew at a very large airport and I can personally say that the time efficiency can't be improved much in any area without a lot of automation and increases in all the bottlenecks at once. The people on the ground need the time the passengers wait to get on board to get the bags from the connecting flights and baggage areas, and the time the passengers board to load them. Even if boarding were finished 10x earlier there would still be a long wait.

All the variables in timing are interdependent.

Feb 15, 2008
I think just charging the plane's ventilation system with a little nitrous oxide before beginning boarding would help matters significantly. :-D
Actually, I think the previous comment carries an important insight: the issue is not time(as it takes all the current available time to prepare the plane, passengers and luggage), but the passengers' attitude about waiting a moment for the person in front of them to complete their boarding. But many people are nervous and anxious about travel, so any additional stress(I guess waiting 4 minutes accounts for stress today) creates a difficult situation.

I figure we're all going to get there at the same time, so chill. And actually, on further consideration, isn't standing in the aisle or jetway considerably more comfortable than the imprisonment of being strapped into that sub-human-sized seat for 3 hours? Enjoy it.

Feb 15, 2008
All good points.
I would add that there is marketing too.
When it is a free for all random boarding (e.g. SW Airlines and Ryan Air) it feels like you are cattle jostling onto the plane. Somehow the inefficient orderly boarding of BA feels more civilized even though it might be slower.
Another point is that you have a very heterogenous mix of passenger types. Business collegues vs a large family group are totally different. Any model needs to include a representative set of passenger types.
I concurr with the previous comment. I don't want to board until the last minute. Sitting in the supposedly average sized person seat is very uncomfortable. Not to mention the fact that you are sitting on top of 1000's gals of volatile AvGas and subjected to a lifetimes worth of pathogens. I would happily pay extra \$10 each trip for the safer military grade JP4.

Feb 15, 2008
Here's a thought: Instead of speeding up the boarding process, why not monetize the time people spend lined up waiting for their seats. Paste up some of those advertising cards used on trolleys and buses - just put them on the overhead storage bin doors. And have friendly commercial messages piped in over the intercom, like they do at gas pumps here in Southern California. They're really quite compelling.
Use the revenue for some of that flower-scented JP4...

Feb 15, 2008
I think if you could simply get the passengers to gutterally chant "go, go, go" in unison as they board the plane, that this would easily speed boarding by an estimated 42%. Maybe more.
I have to say(before we finally move on) that after flying regularly for more than thirty years, on easily many hundreds of flights, I have never felt that I reached my destination any appreciable amount of time later due to how the passengers boarded the plane. Nor do I believe at any time did any of the airlines profits or take-off times suffer appreciably because of passenger-boarding-caused delay. I think we may have just beat, not a dead horse, but a relatively content horse, to death. In point of fact, I think there is actually no problem here, other than the stress(for some) of flying.
But if a vote will settle it, I'm going for the laughing gas, the JP4 and the porters(they're no cost to the airlines as they would work for tips).
Can we now discuss the much more pressing and delay causing problem of getting all the passengers to GET OFF THE PLANE more quickly?!?

Feb 19, 2008
Am I the only one who sees a problem with boarding from anywhere besides the front end?

Feb 19, 2008

Feb 20, 2008
As a frequent flyer of 737s, I have experienced that Southwest flights board much more quickly than other airlines' flights (In my experience, AA is the worst). Southwest does not have assigned seats, so travelers do not have to look for their seats, but can sit in the seat of their choosing -- usually a few rows away from someone who is struggling with their luggage. It would be interesting to add this scenario to the model.

Also, limiting the size of carry-ons to the actual FAA approved size would work wonders.

Feb 24, 2008
Ashibayai said above that speeding up the boarding process can conflict with the onloading of baggage. Studies by Boeing show that in most cases, the baggage is finished before all the passengers enplane. This leaves room for improvement in the boarding process. I don't think Steffen's method is feasible in actual practice, but I do think speeding up the enplaning process would not be circumvented by the baggage process.

Mayday commented saying that he didn't think enplaning kept planes from leaving later. He is correct, but that isn't the point. The point is that if the process were speeded up the planes could leave earlier, freeing up the gate. And after a whole day of flights, there may be a chance to schedule in another flight for that plane, especially if it does short-hauls.