Computer Science 341

`www.cs.amherst.edu/ccm/cs341/`

Computer Science 341: Applied Algorithms, Spring 2012

Course Information

The course instructor is Catherine McGeoch. My office is Room 401, Seeley Mudd Building. My phone number is x7913.
You can send me email at .

Note: I will be away from campus the first two weeks of class due to surgery, with limited email access.
We meet TuTh at 2pm in Seeley Mudd 202. Some class sessions will meet in one of the labs (006, 007) on the basement level of Seeley Mudd.
Office hours: I am available for drop-in conversations MW afternoons (except for the first 2 weeks); or send me a note to set up an appointment for another time.
The course readings primarily consist of research papers that I will provide, and chapters from books that are available via e-reserve and text-reserve in the library. We will cover most of two texts, which you may want to purchase (or share with a friend) to avoid long waits for library access.
- C. McGeoch, A Guide to Experimental Algorithmics, Cambridge Press 2012. Students who order the book directly from Cambridge University Press will can use a special discount code (ask Lyle) to get a 20% discount. This is also the fastest-possible shipping method.
- Z. Michaelwicz and D. Fogel, How to Solve It: Modern Heuristics, Springer 2004.
Grades will be based primarily on two research projects. You will be given a messy real-world problem and asked to: define the abstract problem, implement a suitable algorithm, develop experiments to evaluate it, and write a paper describing your results.
There will also be a few additional small assignments. The second half of the course will be conducted mostly in seminar style: you will be responsible for reading and presentind research papers, and for participating in in-class discussions.
Your work on the first project (coding, progress reports, paper) represents 1/3 of your grade; same for the second project. Your overall contribution in homeworks and course discussions will make up the other 1/3 of your grade.
Here is a syllabus and topic list.
Here is a timeline for the first project.

Course Overview

This course is about understanding -- and shrinking -- the gap between theory and practice in algorithm design and analysis. Theoretical bounds on algorithm performance are important to know, but theory is not always adequate to answer our questions about performance in the real world. This is especially true with algorithms for NP-Hard problems. We will look at how experiment-driven design and analysis can be used to ``flesh out'' the theory, to produce better programs and tighter predictions of their performance. Students will undertake two projects to implement programs to solve real-world problems, and participate in seminar-style discussions of experimental research papers.