**1/12/2006**: Assignment for today's and next week's TD: create the functional and technical software architecture for a program that, given a URL*u*and an integer*n*, draws the graph corresponding to the topological neighbourhood of*u*of radius*n*in the graph*G=(V,A)*where*V*is the set of all HTML web pages on the internet and there is an arc in*A*from page*u*to page*v*if there is a hyperlink in web page*u*referring to web page*v*. Given a web page*u*and a neighbourhood*S*of*u*(containing*u*), the radius of*S*is given by the maximum path length from*u*to all other web pages*v*in*S*.**30/11/2006**: In order to overcome the limits on problem size in AMPL, perform the following steps:- in all your AMPL
run files, replace the string*option solver cplexstudent;*with*option solver "/users/profs/info/liberti/bin/cplex"*; - instead of calling AMPL with the command
*cat amplfile.mod | ampl*, use the command*cat amplfile.mod | /users/profs/info/liberti/bin/ampl*.

- in all your AMPL
**15/11/2006**: I updated the AMPL utilities and the notes to include the automatic determination of the number of clusters. Please also take a look at the updated exercise book which contains the solution to the functional architecture. Try it yourself, change the AMPL model and data files, see what other types of functional architecture you can come up with.**TD of 10/11/2006**. Homework: download the AMPL utilities for the reformulation of the functional architecture graphs. Some more information about AMPL, including how to download it, in software section of my operations research course. If all else fails, use AMPL in the computer labs. Get reasonably comfortable with it: use it, change the models, look for new models on the internet, implement them; relax or soften constraints from the existing models, see how clustering or interfacing change when you change the constraints coefficients or right hand sides.**Projects**. Start thinking about a project to be completed for examination purposes. Here is an example of what a software architecture description might look like (keep in mind this is just a very short example for didactical purposes only, so it is neither complete nor sufficiently detailed; your effort should be better than this!).**TD of 27/10/2006**. Homework: exercises 3.1 and 3.2. Use this DBDesigner file to form the SQL queries which answers the questions in exercise 3.1. Start thinking about initial and final quantitative indices to cluster the projects, and which data (and SQL requests) we shall need from T-Sale's database. Read up the notes and the two papers on clustering listed in the didactic material section (send me an email when you finished downloading them!).

- Lecturer: Daniel Krob
- Assistant: Leo Liberti

- Course website
- Exercises
- Notes on quantitative software architecture techniques
- Review on data clustering
- Data clustering and mathematical programming paper