Computer Science 858 (Detailed Information)

Meeting Times

Tuesday, Thursday at 10:00-11:25am in Thorvaldson S386.

Instructors

Class Instructor: Nathaniel Osgood, Thorvaldson 280.6, 966-6102, osgood@cs.usask.ca
Office Hours: Tuesdays 3:30 – 5:30pm and by appointment.

Texts

Readings from a variety of sources (primarily articles and excerpts from texts on dynamic modeling) will be provided as PDFs on the course website.

Brief Focus

Introduces the theory and practice of Agent-Based simulation modeling for informing decision making and strategic planning, with a strong emphasis on applications in public health.

Course Contents

This course introduces students to the theory and practice of agent-based simulation to inform decision making and policy evaluation, particularly in public health. The course provides an introduction to the use of simulation using examples drawn from the health area, and covers the entire model lifecycle from qualitative model conceptualization, quantitative model conceptualization and formulation, model parameterization & calibration, use of models for policy analysis. This course is highly complementary with – but does not require as a prerequisite – the version of CMPT covering System Dynamics modeling.

While public health simulation models have been proposed and conducted in a variety of frameworks the course will focus on modeling using Agent-Based modeling. Students will be provided with some additional background in order to come up to speed in each of these areas. While System Dynamics modeling is taught in an alternating version of this course, the course will provide a very brief exposure to System Dynamics modeling tools for interested students.

In light of the diverse background of course participants, the instructor will strive to present material at a level that will be accessible to all participants, and extra tutorials on background material will be presented upon expression of interest by students. Class exercises and assignments will be carried out in user-friendly graphical software. However, in contrast to the System Dynamics version of the course, some computer programming will be required to complete course assignments. Following the tradition of CMPT graduate “topics” courses, this class will review pieces of the associated literature in a more informal manner. While the class will include presentation of material, most such delivery will be somewhat informal, and the “instructor” will often serve as more of a facilitator with session time will be devoted to hands-on in-class exercises and tutorials aimed at fostering learning for all involved. Graduate students within the class are assumed to be mature colleagues that will participate in active learning by identifying and seeking out material of interest. In the course of the semester, the flow and direction of the class may be adjusted to reflect the interests of these participants. The schedule included in this outline is thus tentative.

Grading

The weighting for coursework is as follows:

Note that a significant fraction of students’ grades will be based on class participation. In recognition of differences in communication styles and interests among students, this participation score will reflect interaction in class and in office hours and (where appropriate) tutorials.

The contents of the term project will be decided under discussion with the instructor.

Project

The completion of a project is an important part of the class. The project type and topic will be chosen by the students in conjunction with the faculty, and may be pursued on an individual or group basis. Students are encouraged to work in teams for the project. Based verbal discussion and on a series of incremental deliverables submitted during the term, the instructor (and possibly individuals) will help guide the students on the project. (Please note that these early deliverables for the project may be informal and rough. While all deliverables should attempt to communicate material, no focus on creating a careful presentation of these materials is required prior to the final report for the project.) The project emphases may vary, but most projects should include the following:

1. Creating one or more simulation models that describe these phenomena, or adapting a particular model to a different context.
2. Placing data into the model to customize it to a particular context.
3. Running a “baseline” scenario with the existing model parameters, and commenting on its plausibility.
4. Running one or more “what if” scenarios with the model to explore different possible situations. These situations could reflect the results of implementing different policies, or different possible external conditions.
5. Performing one or more “sensitivity analyses”, in which assumptions in the model (in the form of parameter values, or structural elements of the model) are changed.
6. A well-structured written report describing the above, and your findings.
7. An in-class 30-minute presentation where you summarize the model, its construction and findings.

Please note that while most projects will fall into the framework described above, the instructor is open to the idea of students exploring alternative methodologically oriented project types. These could include (but are not limited to) investigation of meta-level issues in modeling (e.g. challenges in characterizing a certain dynamic phenomenon in different modeling frameworks), or an analysis of or proposal for innovations involving the modeling tools and methodologies used for simulation modeling. If you wish to pursue one of these other types of projects, please discuss your interests with the instructor.

Course Website

Students may access the course website

We will try to post course lecture slides, notes, and other material to the website.

We plan to record the lectures given as part of class, for reference by students in this class as well as others. We are asking for student help in recording these lectures, and will attempt to post them to the course website on a regular basis.

Course Reference Book

There is no required reference book for the class. Interested students may find the following two books of value in understand elements of Agent-Based Modeling and System Dynamics modeling, respectively:

Epstein, J. M. (2006). Generative social science: studies in agent-based computational modeling. Princeton: Princeton University Press.

Sterman, J. Business Dynamics. Boston: McGraw-Hill Higher Education. 2000.

Prerequisites

This course will demand more than is typical of a class. This course will require some dedicated effort from each student, but that effort will be met with dedication on the part of the instructor. The demands are of several sorts:

The class will have a significant time demand, and we do not recommend that the class be taken for credit by students whose schedules prevent devoting many hours a week to master course material. However, much of the demand of the course will come not in terms of time, but in terms of willingness to stretch yourselves conceptually and in terms of your knowledge base. While the class is highly interdisciplinary (probably more so than any class the students have taken in years), it will demand much “stretching” from all sides. What is being asked of students will vary by area of the students:

• For those from health sciences background: A willingness to take on quantitative challenges, and to acquire new skills and approaches to understanding public health phenomena. For this year in particular, the course will require a willingness to learn small bits of computer programming.
• For those from Computer Science background: Patience with understanding textured health science concepts, terminology and aspects of public health practice.

All types of students are also being asked to show respect for the skills and perspective brought by other students, regardless of their area.

A serious investment in the above will give you a set of skills that are currently rare and much valued. However, based on the level of commitment being required, we recognize that this course will be not be a good match for many students, and recommend that students who are not confident of their wiliness to commit their time and effort to the course look elsewhere.

Tentative Topic Plan

A preliminary schedule for topics is included below. As noted above, this schedule is tentative is subject to change as student interest and discussion direction warrant.

NB: The schedule above – including the contents and dates of both lectures and management meetings – is subject to change. On occasion, updated schedules may be provided.

Where these occur, lectures slides will be provided via the course website (using WebCT). Readings distributed electronically are shown below. Additional readings and URLs will be shown in the notes for the topic in question.

Tutorials & Labs

The class will include a number of tutorials and labs to provide background material for the mathematical concepts and software used in class, or to provide more advanced material for those students who have the requisite background. These tutorials are optional, and will be scheduled per the availability of students. To facilitate meeting as a group, we will attempt to arrange a specific tutorial time to use for the entire term.

In some cases, a possible tutorial topics may instead be presented as a lecture.

Possible areas for background tutorials are as follows: