Very cool. It's interesting to note that Elevator Dispatch Algorithms are very similar to harddisk read/write algorithms. Hard drives have to accept read ("pick up") and write ("drop off") requests at different disk addresses ("floors") and figure out the most efficient way to handle all of the requests. In this analogy, traditional hard disks are simply very, very tall buildings wrapped around a cylinder.
My dad is an elevator mechanic and just recently switched from Schindler to Thyssen Krupp. It's always boring to go see him at work now - all the new buildings as well as the modernized old buildings are almost exclusively switched over to digital systems. The old systems are incredibly interesting - huge cabinets full of relays! The smell of ozone and the sound of thousands of relays clicking in consortium is really enjoyable to see. Imagine the relay computer videos that come across here on occasion, but put in ten large cabinets of the relays and you get an idea of the sound.
Of course, touch the relay without a relay pick and you'll get a 120v shock :)
The stuff put together in the 80's is a combo of relay logic and old school processors, where you get to load the program in using paper tape!
Do you know of any videos where we can hear the sound of that many relays? I believe i've heard sounds similar to what you are describing in movies, but I'm not sure if those were the same thing, or if they were accurate.
When I was a kid my father worked for Southern Bell. I used to go in with him on trouble calls to the local office and it was a building full of rows and rows of relays like those. Smelled awesome.
What's funny is just how simple these algorithms are. All it really took was something looking at this from a slightly different perspective.
The best optimizations I've done at my job have been exactly that. Looking at things a little differently and even removing steps. It's really a testament to open-mindedness and creativity.
Here is my favourite elevator optimisation story [1]. Customers were complaining that the lifts were taking too long to arrive. It turned out the root problem was boredom, and the solution was to install mirrors in the lobby. The mirrors allowed people to fill in the time by admiring themselves and others. A neat illustration of how there are sometimes simple non-engineering solutions!
---
Update: Q: Have I been suckered by an urban myth here? I have a vague recollection of knowing this story before the publication date of the source I gave, and it seems to be well used in "management" books but no mention of specifics or a primary source.
Further poking around comes up with an article that claims the story originates with a consultant friend of Russell Ackoff (the author of the book I linked to), and took place in a New York hotel in 1955. A bit more specific, but I'm still a bit suspect since Ackoff doesn't seem to mention these details in his writing. Presumably if it is mentioned in the earlier book that you point to, Ackoff must have published or mentioned it elsewhere. I've done this one to death now, so I think I'll move on.
Basically some problems are made trivially easy, without having to change the problem definition by changing the underlying representation of the problem.
So how long before they can predict the future to be at the right place at the right time every time?
It's certainly how interesting how critical a good setup can be to get a large skyscraper to actually work well. If they take too long to get to the people waiting then they need to leave or get to work. The old game SimTower is actually a pretty decent simulation of this kind of effect (if not still overly simplified).
SimTower was /actually/ just built around an elevator simulator. "You guessed right," she said. "Sim Tower was built around a real elevator simulation program we bought from a Japanese guy." http://www.gamasutra.com/view/feature/2150/the_designers_not...
I was wondering when someone would crack a Hitch-hiker's Guide reference out. Well played.
More substantially, I agree with your comment - it's going to be money invested in the pursuit of giving humans more time to do the things in life we care about, and that's a Good Thing.
In the early 90s I did a high school science project at a research lab that was working on a smart elevator project for Otis. Everyone had a NeXT workstation and a Symbolics LISP machine. And they wore ties. It's hard to imagine getting a job now where you would get that much money spent on your computers. Most people are lucky if they get two screens.
Actually I'm surprised by the praise. This kind of elevator isn't uncommon here, both in towers and hotels. And it makes so much sense..
In both scenarios you probably want some kind of access restrictions, be it 'You work on floor 10 so that's where you can go" or "You can go to the cafeteria on the top or the floor of your room".
This system, the way I know it, solves both this problem and allows the elevator software to plan ahead.
Side note: Paternoster are quite cool. I prefer them to most 10-20 year old' real' elevators.
the Miconic 10, a smart elevator system that allows users to enter their floor onto a keypad.
"The first U.S. installation of Miconic 10 was at the Ameritech building in Indianapolis in 1993. The next generation added a radio frequency identification, or RFID, reader to the keypad to identify passengers, so tenants could swipe their ID badge or card to call an elevator. In 2009, Schindler launched the third generation, called PORT (it stands for “Personal Occupant Requirement Terminal”), a software-driven system that uses a touch-screen and learns movement patterns, enabling it to personalize its offerings to each passenger.
I was just at a hotel where I had to:
1. press the up button
2. swipe a room (key) card on the elevator
3. often swipe again because the card needs to swipe a certain way, swipe for other people who were equally confused
4. press the floor number(s)
versus
1. press the up button
2. press the floor number
I would imagine that guests would have a room (key) card with RFID, and that by default the lobby elevators would read the card and by default offer to go to the floor that the room is on, and from the room floor elevators it would by default offer to go to the lobby. The "keypad" or touchscreen would only need to be used if you wanted to go to another floor: pool, restaurant, conference room, etc...
The elevators need to clearly indicate where they are going.
At work:Six elevators, two (?) panels per floor.
If you come near one of these panels it lights up and offers a touchscreen with all possible floors. By default the lobby and the parking deck.
If you swipe your access card in front of that panel, the selection is expanded to include the floors you're cleared for.
Select a floor and you'll be told which elevator door is going to open.
So: For leaving the building: One press. For your way up to work. Swipe, press.
Up and down buttons are idiotic if you specify the floor anywhere..
