General:

Final assessment is by a  3,500 word term paper (100 %) based on a suitable programming project.

What's expected:

In order to gain a good mark we expect more than just a working implementation. a project with just a working GA, neural network or control system may still merit a fail mark. A comparison of different techniques, different parameter settings, encoding schemes etc. or some other kind of exploration of the problem and the solution techniques. The write-up should include a clear description of the methods you have used, an algorithmic description of your implementation, a clear presentation and analysis of results including systematic collection of statistics and control experiments, and commented program listings. You must draw some significant conclusion from your study. Creative exploration of novel aspects is strongly encouraged. Comments at the end of the options below give you an idea of which are more difficult and would therefore require a less exploration to gain a high mark. Overall difficulty is a conserved quantity, so more work is expected from you if you choose an easier project.

Format:

Please, write the report clearly specifying the objectives, background, method, results, and implications of your work (exactly as you would with a research paper). Divide it into sections. You must include enough contextual information about the background of your projects, including, when suitable a short critical review of relevant work. Try to confine any discussion related to the code, implementation problems, etc. to an appendix. Include a well-commented version of the code as another appendix.
 
Suggestions:

For suggested projects, see here.  Code for the path search problem is available here. The file low-inertia-robot.pdf  (Option 3) is available here