CS 2390 Winter `95

DUE MARCH 7 IN CLASS

A Simulation of a Vending Machine

Using the simulation package we've been discussing in class (in ~mg65/2390/SimulationPkg.st), create a simulation of a Vending Machine which contains three kinds of soda: Coca-Cola, Mountain Dew, and Sprite. The potential buyers arrive on a uniform distribution every 1 to 2 hours (Uniform from: 1 to: 2). They are twice as likely to purchase Coca-Cola as they are to purchase Mountain Dew, Sprite, or nothing at all (i.e., not thirsty). They are equally likely to purchase Mountain Dew, Sprite, or not order anything. If they're not thirsty or can't get what they want, they simply walk away. (N.B., this isn't the default behavior for the Simulation Package!)

The soda delivery person shows up on a uniform distribution every 48 to 72 hours. Here's the trick: The Vending Machine sits on top of Stone Mountain.

• The delivery person has to hike up the mountain carrying the cans. He isn't very strong, so he can't carry more than 15 cans of soda on a trip.
• He's also a persnickety fellow: he refuses to carry the exact same load everytime up the mountain, so there must be some variance in the load (e.g., not a constant number of each kind of soda everytime up the mountain.)

Your job is to come up with a strategy that maximizes the ratio of Satisfied buyers (i.e., buyers who walk up to the Vending Machine, do want a soda, and are able to get the soda that they want) to Unsatisfied buyers (i.e., buyers who walk up to the Vending Machine, do want a soda, but can't get the one they want because it's not there.) in a 200 hour period. You basically can fiddle with the make-up of the sodas that the Delivery person delivers (more of one kind, less of another).

In your writeup for this project, hand in 1-2 pages (1) describing at least TWO other options (besides the one you're handing in) that you tried and (2) evidence of how well your final strategy works from at least TWO runs (you decide on the kind of evidence that best proves that your strategy works, e.g., histograms, statistics, etc.) Be sure to indicate the commands that the TA needs to execute to run your simulation.