Optimization is getting the most out of a situation with fixed constraints.  In our building of 10 floors with 4 elevators we have 800 or so employees.  As with most office buildings, we all arrived around the same time and leave around the same time.

The optimization statement is:

Goal: Minimize wait time for elevators, after the ‘call elevator’ button is pressed, therefore increasing worker efficiency.  Wait time for the elevator can also be communicated at total distance traveled for an idle elevator to a travel to a floor that it is called to.

Measurement: Travel distance for closest idle elevator to travel to a floor where the ‘call elevator’ button pressed.

Constraints: Four elevators, nine floors, all elevators must service all floors (no dedicated elevators).

Variables available to change: Location during idle time of the elevators (meaning: where do the elevators hang out when no one is in them).

Time of day and usage relationships:

Every morning people come in on the first floor, and go up to the other nine floors.  This is a one-to-many relationship.  In the middle of the day, it turns to a many-to-many relationship, as people travel from floor to floor.  And, at the end of the day the demand shifts again to a many-to-one relationship, as people go home.  You can also assume a shift again near the lunch hours, we’ll return to this later.

The idle elevators should move to a position, automatically, which will best serve the usage relationships through the day.  This would look like this:

Morning/post lunch: One-to-many relationship – All elevators return to the 1st floor automatically when idle.

Mid morning/mid afternoon: Many-to-many relationship – Idle elevators return to floors 1, 4, 7, and 10.  This pattern puts an elevator within a floor of all floors at all times, while leaves an elevator ready for ground floor arriving guests, and prioritizes an elevator for top floor executives.  The 1,4,7,10 relationship is maintained as elevators are used, so empty elevators move to keep the balance as in-use elevators deliver riders to their chosen floor.

Late afternoon: Many-to-one relationship – Idle elevators return to floors 5, 7, 9 and 10.  Elevators will rarely be idle during peak times, but it should be reinforced that most elevator use will be taking people down rather than up.  Therefore, moving the idle elevators up to higher floors makes them better positioned to carry people down to the first floor.  It can also be assumed that the lowest floors use the stairs at the end of the day already because their position nearest the first floor will always result in an already full elevator being delivered to them when pressing the ‘call’ button.

And that’s it.  This would take some programming of the elevator’s master control, but probably not much.  The key take away from this is to minimize the distance that an empty elevator is traveling, reducing waste of ‘time’.  Why did I think about this so much?  Well…let’s just say I find myself standing waiting for the elevator often.

Post filed under Other Ideas.