csci 8314

C S C I     8 3 1 4

Announcements

• May 9th, 2019 : Regarding the grading of the projects: I am grading with the following criteria:
```	1. Clarity and quality of general presentation: 30%
2. Relevance to course: 20%
3. Originality: 25%
4. Amount of work done for the project [any coding? quality of references, topics covered
and related detail of description..] 25%
```

• Apr 29, 2019 : Due to an exam, I may be late for my Tuesday Apr. 30th office hour. I will be in for sure from 2:00 to 3:00pm. If you can't make it please send me an e-mail to schedule a meeting at another time.

• Apr 12, 2019 : Regarding HW2.. There were a few questions regarding the matrix A produced by the get_mesh script. This is not a graph Laplacean and so you will need to create a Graph Laplacean from it. You will need to do this inside your RSB1.m script - because the driver assumes it is done this way. I mentioned this last time in class but I dicovered that the HW statement was not clear about this. [Now corrected].

• Mar 19, 2019 : I graded HW1 -- and will be posting the grades tomorrow. There were a few quite common mistakes:
• Confusion between fill-path and fill-edge. A path is a sequence of vertices, like 4, 3, 2, 1, 5 for example. An edge is a sequence of 2 vertices: 2,5 for example In Exercise~1 you were asked for fill-paths in order to show what fill-ins will take place. Many (if not most) just showed the fill-ins of Gaussian elimination [or the fill-edges]. I took off 2 pts for this and another 1 pt if the error is repeated later (in part (e) - after the RCM ordering).
• In the GP algorithm: many of you did not set the vector properly. If you say
```	    [ii jj v] = find(A(:,j));
Lst = TopSort(L, ii);
w = sparse(ii,jj,...)
... or ...
w = sparse(n,1) ..
... or .. many other similat statements
```
then there will be fill-in -- meaning that internally matlab will reallocate to accomodate the new entries. You do all the work with TopSort - precisely to avoid this! I took off 4 or 5 points (depending on the situation).
• For Question 3 - part (a) - many of you gave a much more complex counter example than needed. I usually did not take off points for this..
• Finally, for Question~2 part (a), a few of you used the reverse permutation to get the adjacency graph. When you see pi = [1, 6, 5, 3, 2, 4], it means the numbers 1, 6, 5, ... are the old labels and the new graph will relabel these as 1, 2, 3, ..., (new labels).

• Mar 09, 2019 : I updated the Project by adding references. This file will occasionally by updated.

• Mar 06, 2019 : HW1 is due on March 13 (not 15th as posted which is a Friday). However there will be no penalty if you turn it late until the 15th at noon.

• Mar 05, 2019 : I posted some scripts that you will need for HW1 (q. 5) -- see the matlab page.

• Feb 6, 2019 : Posted a practice exercise for class of Feb. 11th.

• Feb 6, 2019 : Please note that my office hours are Mo-Tu 1:30 -- 2:30. [Earlier posting was incorrect]

• Jan~2, 2019 :
Welcome to csci 8314 !