I've written a basic elevator and a complete simulator to go along with it. In the simulation, people are randomly generated on each floor and each have a random floor to go to. The elevator tries to minimize the average wait time, while also bounding the maximum wait time.
I've tested this on Python 2.7. To run,
python simulation.py --floors 50 --iterations 100 --max_new_people 3
A detailed description of each command line flag is given by
python simulation.py -h
You should see command line output like
Mean waiting for elevator time: 10.806
Mean waiting on elevator time : 7.333
Mean total wait time : 18.139
If you have matplotlib installed, you should also see some plots giving more detailed statistics. For example,
simulation.py contains all of the simulation code, while elevator.py has the elevator business logic.
(sorry about the 8 space tabs - it's 4 spaces locally)