This week, things moved pretty fast. Minh Tu learned the algorithm. Both of us still have difficulties on finding the value for (alpha). We have not gotten the concept sinked in. Other than that, we got over the main hurdle - coding out Gauss-Seidel Algo for sparse-matrix in PHP.
Dr Chong will now verify it. I have tried using sample matrices found online and the results for it are a perfect match with the correct answers. An important reference for us was :http://numericalmethods.eng.usf.edu
So here is how we started(and ended) our coding session.
Monday - we met Dr Chong and Minh Tu was made clear of the algo.
Tuesday - we laid our the pseudo-code and decided to call it a day
Wednesday - The algo was coded in php
Thursday - We did some testing but the problem was all the arrays given were non-sparse so we modififed the code a bit to suit for non-sparse matrices. Tested out a few sample matrices found online and the resultant vector is a perfect match.
Friday - We had a figure our a way to stop our looping. Right now we had not put in the epsilon function but instead had hardcoded it to loop 100 times. Only thing left is to use epsilon. Nothing much was done today.
So that is how we completed the code for Gauss-Seidel Algo. Now our main focus is to do the peripheral stuff like coding for facebook(which is actually the harder part for us)
Friday, May 22, 2009
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment