Assignment 5: Report and Marking Scheme

Report

Details will be available by Monday.

For your report, you are to produce a single pdf, with 2.5cm margins, two columns, and 10pt font (Times New Roman). Aim for half a page of content. Max two pages including citatons. This should include your and your partners' names and student numbers. Add the pdf, named report5.pdf, into your "a5" repository, and commit the file. Be sure to cite any sources you use.

I recommend LaTeX for producing your documents; you may, however, use Word or OpenOffice and convert the file to pdf. A single bonus mark will be given to students who use LaTeX to produce their reports. If you do so, hand in the .tex file along with the .pdf file to demonstrate to the marker that you used LaTeX. Feel free to use the template from report 2: LaTeX template for you to use (output).

There are three sections to your report; please indicate them clearly.

• Part One: In this assignment, we modelled STIs using graphs. This is only one way that epidemiologists will model the spread of disease. Models using ordinary differential equations are also commonly-used. See the pdf at Simple Epidemic Models to read about the SIS model there.
  1. In the SIS model, dS/dt = bI - aSI and dI/dt = aSI - bI, where a is a constant called the infection rate, and b is a constant called the recovery rate. In Part One of Lab 10 (our SIS model), we had two constants as well: the probability a person will transmit the disease (p), and the time it takes a person to recover (time_to_recovery). How can we relate our two constants to those in the ODE models?
  2. SIS models have two steady states: one where the disease blows over after some time, and one where it becomes endemic. From our two constants (p and time_to_recovery) and the size of our population, how can you predict which steady state a simulation will arrive at?
  3. Consider Part Two of lab 10: instead of becoming susceptible after some time, people die. What would dS/dt and dI/dt become in this scenario? How else would the ODE-model change?
     
     
• Part Two: Consider these simulations from test_bacterial (see test_vaccination_extended.c):
  • Simulation A: n=16, p=1, time to recovery (ttr)=10 (medium)
  • Simulation B: n=64, p=0.05, ttr=6 (large)
  • Simulation C: n=512, p=0.5, ttr=5 (huge)
  For each of these simulations:
  1. Determine R0, the basic reproduction number. R0 = (a/b)*n.
  2. Determine V, the percentage of people we need to vaccinate in order to ensure the steady state will not be an epidemic. V = 1 - (1/R0).
  3. How does V compare to what we get from vaccinate_randomly for each of those five populations?
     
     
• Part Three: Thinking beyond our models
  1. In all our models of viral STIs, we ignored viral load: an infected person's infectivity is not constant over time. In reality, it decreases with hime, and then increases with subsequent reinfections. (And in the case of HIV/AIDS, increases when a person goes from HIV to AIDS, as their immune system is destroyed.) How do you think adding viral load to our models would change them?
  2. All our models have been for STIs. How would our models change if we were modelling non-sexually transmitted diseases? How would you model measles and other airborne diseases? What about epidemics that are not necissarily spread human contact, such as salmonella?
  3. How would the route of transmision (sexual, airborne, etc) affect the number of people we need to vaccinate? Why is it so important that we need to vaccinate everybody for diseases like measles and pertussis?
  
  
  

Marking Scheme

• Assignment 5 will be marked out of 100.
• 20 points -- your code demonstration in Lab 9
• 10 points for code style
• 1 point for having run gprof; 1 point for having a Makefile
• 3 points for completeness (1 for each of Parts 1, 2, 3)
• 5 points for your ability to explain your code
• 20 points -- your code demonstration in Lab 10
• 10 points for code style
• 4 points for completeness (1 for each of Parts 1, 2, 3, 4)
• 1 point for effort
• 5 points for your ability to explain your code
• 30 points -- automarking of your final versions of the Lab 9+10 code (10 bonus marks available)
• 2 points for your Makefile
• 14 points for lab 9:
• -1 for any memory leaks, compiler warnings in your code
• 2 points: optimization in Part Three
• 1 point: correct output for Part Three
• 5 points: Part One correctness
• 6 points: Part Two correctness
• 14 points for lab 10:
• -1 for any memory leaks, compiler warnings in your code
• 3 points: Part One correctness
• 3 points: Part Two correctness
• 3 points: Part Three correctness
• 2 points: Part Three efficiency
• 3 points: Part Four correctness
• max 10 bonus marks for Part Five
• 30 points -- your report (1 bonus mark available for using LaTeX)
• 10 points for each of the 3 parts:
• 1 point for addressing each of the 3 subquestions
• 5 points for good reasoning
• 3 points for clarity of writing; spelling, grammar, etc.
• 1 point for citing sources where appropriate
• 5 points will be deducted if the three parts are not clearly distinguishable (they are marked by three different TAs!)
• 1 point will be deducted if your names and student numbers are absent, or your report is not named report5.pdf (this stacks)

CSC 190

Computer Algorithms and Data Structures

Navigate:

• Information
• Schedule and Lectures
• Labs and Assignments
• Code Style
• Exams
• Other resources
  • Piazza (for course questions)
  • Blackboard (used only for grade posting)

© 2012 Elizabeth Patitsas
Template by Andreas Viklund