• Add a Comment
  • Edit
  • More Actions v
  • Quarantine this Entry

Comments (1)

1 thecolour commented Permalink

They have these things in my building (Shell Center, Waterloo, London).

 
They suffer from one major drawback, but I think this drawback makes this model of elevator impractical for many deployment scenarios.
 
The drawback is that the elevator cannot know the number of people waiting for it and has to guess, based on the number of requests for a given floor.
 
This means that if you make multiple requests for a floor, your chances of getting an elevator sooner are actually better (as the controller will route a greater number of elevators to your floor), but this works to the detriment of the general efficiency of the elevator system as a whole.
 
To put it another way, individuals are incentivised to "button mash" (ie, make multiple requests) because this makes sense from an individual point of view. However, this behavior means that while a few individual journeys are quicker, most journeys are slower as the system "punishes" users who tow the line.
 
Furthermore, this effect is greatest at times of highest volume of traffic. In other words, the busier the system, the greater the reward for deviation and the greater the "punishment" for being nice.
 
Any thoughts from others?
 
--V interesting topic I think. I am fascinated by the optimisation of destination programming for elevators..

Add a Comment Add a Comment