How to speed up elevators with an algorithm

Today I want to tell you about an idea I had some weeks ago and how I try to make this an open innovation.

I came up with an algorithm that could increase the speed of elevators.

Half of the time using an elevator is approximately used to order it and to wait for it in front of the door.

Why is this so? And why doesn’t the elevator wait where I am?

Ok sounds like a silly question how should the elevator know where the next person comes. He can’t, at least not easy. You could do something with a bunch of sensors who track if someone is coming near, getting slower and than place a elevator in front of him just in case. This wouldn’t work in my companys building. Elevator is next to the toilet and it would have to work very good, that it doesn’t move up and down all the time.

But luckily I’m a computer scientist specialising in data mining and prediction is also my favourite topic. So why not use the knowledge I have to improve elevators.

I would love to but for data science you need data. If anybody here could provide me with data, I would be happy. When do elevators stop where?

So I came up with a more abstract idea, I could just guess data. I assume every building has a, I called it, heart beat. This heart beat is when people come in and people go out.

I made this example for my office building (6 floors) and I think it will work for most offices and maybe also for other buildings (german counting, 0 = ground floor) .

There is a time in the morning where everybody shows up at work, this is between 6 and 10. During this time nearly every ride is from ground floor to one of the upper floors. People in floor 1,2 and 3 are more likely to walk. So

So during this time the most likely ride is between 0 and 4, 5 or 6.

And I think it would be faster if the elevator is not used it could instead just staying where the last stop was going back to the ground floor and wait there, because chances are high at this time that someone is coming instead of leaving.

For lunch break most of the people are leaving the house or going to the cafeteria, which is also in ground floor. So guess what between 11:30 and 12:30 most of the people go from the upper levels to the ground floor.
And there is also the same logic for the evening hours.

But every building is different. Some start working earlier, some have parking spots in the basement. Some have the cafeteria at the top floor. The heart beat could also change so a hard programmed or fixed time frames wouldn’t work.

The elevator is improving the customer experience during his free time.

As I already showed there is a rhythm, so it is possible to make a forecast. Maybe the pattern is in some cases more complex but should be easy enough for a algorithm to recognized. And that is where the cool stuff comes in. Here you could very easily implement a self-learning algorithm to predict where it is most likely that the next passenger comes in and the elevator could use the time he is empty and just waiting to improve the customer experience for the next guest.

I already have two different algorithms in mind, but no data science without data.

And here is where the open innovation comes in: If you are interested in this idea feel free to contact me. I'm also very interested in getting data to prove that my idea actually works and how much time could be saved.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.