Due: Thursday, March 2, 2017, 8:10 am (start of the class)

## Problem:

Write a simple discrete-event simulator of an M/M/1 queuing system with a bounded queue (such systems are referred to as M/M/1/k, where k is the size of the queue).

- To test the simulator, design and run experiments to verify that simulation results are in line with known analytical results for the case where the queue is unbounded. It is sufficient to focus on just one measure, for example the mean waiting time.
- Experiment with finite-size queues and show the impact of different queue sizes on the performance measure studied in the first part.

The main goal of the assignment is to evaluate the correctness of simulated results by comparing them to theoretical results. The simulator itself is a tool but not the main outcome of the assignment. The design and execution of the experiments constitute a significant portion of the assignment. A perfectly working simulator but no results will earn you only a fraction of points. It is expected that you will be making choices and decisions on your own as you design the simulator and the experiments. Describe the choices in your report. If you are unsure, ask.

## Deliverables:

Write a **report** that includes:

- a brief description of the simulator
- a detailed description the experimental evaluation (there has to be enough details so that one can reproduce the results)
- results of the experiments - including a graph that compares the theoretical results with the observed ones
- a brief conclusion

The **source code** is to be submitted using the git repository that you created for the previous assignment.

## Submission instructions:

Upload the paper in the PDF format using Canvas (mycourses.unh.edu) and push your code to the repository. More details can be found in the standard assignment submission instructions.

© 2017 Radim Bartos.