Basic HTML Version
© 2014 Pacific Crest
439
Scenario:
How many doctors should a hospital have on duty to provide service for all patients
in a reasonable amount of time?
Step
Watch it Work!
1.
Do you need a
simulation?
Number of doctors needed in a hospital. An experiment may cost
lives.
2.
Choose the probability
distributions for the
input(s)
The distribution of how many
patients are at the ER is given
by computer records for the
hospital.
Patients
Probability
0
0.15
1
0.25
2
0.30
3
0.15
4
0.10
5
0.05
3.
How do you sample the
inputs?
Random numbers are chosen
between 0 and 1. The numbers
are divided according to the
distribution of probabilities.
Patients Random Number
0
0 – 0.149
1
0.15 – 0.399
2
0.40 – 0.699
3
0.70 – 0.849
4
0.85 – 0.949
5
0.95 – 1
4.
Run the simulation
The simulation is run a million times with the current number of
doctors on staff
5.
What is the distribution of
the output?
The output distribution reveals the number of patients waiting at any
given time.
6.
Vary the input distribution
and repeat steps 4 and 5
The number of doctors on staff is increased and the simulation run
again. The distribution for the nights and weekends is also used.
7.
What is the overall pattern
of the outputs?
The simulations reveal that adding 1 more doctor would significantly
lower the chance that two or more patients be waiting at any given
time.
8.
Validate the results
The hospital staffs another doctor and records the wait times. The
results agree with the simulation.
YOUR
TURN!
Scenario:
An investment opportunity depends heavily upon the strength of the economy.
Suppose that an investment has a 20% chance of resulting in a profit of
$50,000, a 70% chance of a profit of $10,000, and a 10% probability of a loss
of $60,000. Run a simulation and determine whether the investment is a good
idea or not.
9.4 Simulation