Source. There's a summary in Russian.

Basically, the Japanese were experimenting with getting people out of a building as fast as possible. They discovered that putting a column to the side of an exit - not in front of it - funnels the crowd out the doors faster. Otherwise, people approaching the exit all attempt to occupy the same spot in time-space and end up getting stuck... in true sheep fashion. :P

Edit: Wait, I was going for Natural Disasters.

OK. It took a lot of reading and my maths are not great.
I found the sections VII and VIII where they did verify that it was a real effect.
I am still not sure about the values they used for friction and turning effort on pedestrians.

I don't read Russian so I assume the summary is "Put some kind of obstacle somewhere on one side of an exit if you need to move a lot of people through it."

So what kind of effect would placing the exit near a corner of the room or close to a wall have?

I have copied and I will read again.
I am not sure if I can use it but it is interesting.

I went to a lecture about this nearly 20years ago, my university's computer science dept worked on this stuff after a fire at a tube station in London in 1988.

It's surprisingly counter-intuitive how people move. For instance when they come out of a narrow corridor into a large space they stop - it's safer to put exits at the end of corridors. You also don't want to put them in the corners.
The original tests involved ants and plastic mazes

The most important conclusion was that people move with the air flow = toward clear air. This is what caused the most casualties in the tube fire, people went up the stairs - but because of a chimney effect so did the fire.

There's a book called "Simplexity" also a theory by that same name, that states things that are seemingly simple are made very complex, ie. the cell phone and the 65 page user's manual...and complex things are somewhat simple, as in crowd control. There are columns and obstacles at exits, by elevators, and such locations for a number of reasons. One of which is that the few seconds it may take the crowd to move around the obstalce also helps move the crowd which has already moved past the obstacle. The crowd behind is slowed by the obstacle, thereby reducing the chances of too many people occupying too little room. Very fascinating read.

I read it again.

They need bigger exits and they need something resembling a travel corridor to conduct people away from the exit after they get through it.

They have only considered what happens inside the exit and have no accounting for the back pressure caused by people hesitating or running into obstacles as they leave the exit.

Just found these pdf files from UK government research. Someone may find them interesting

http://www.cabinetoffice.gov.uk/epcollege/news/understanding_crowd_behaviours.aspx

Thank you PureSurvival.
It looks quite interesting to me.
I was just skimming it and found more than a few paragraphs that are definitely worth a close read.