My implementation of Chris Walshaw's Multilevel Algorithm for Graph Drawing works! Actually it worked last Friday (19/11/10) but I wanted to test and write a mini-report to ensure it wasn't some fluke.
I have successfully received useful output that can be drawn, with output looking similar (but not the same as) that of my Peter Eades Heuristic for Graph Drawing Algorithm. There are a few problems with my implementation such as no soft-coded cooling function and the level of coarsening depends on how many times the user calls the method, but all can be fixed given time.
In completion of this, I have written a small report detailing the algorithm and my implementation, so 1. I have something more detailed that "it works" when I come back to it, and 2. in case anyone needed it. Output has still yet to be recorded, albeit I have been doing this for today (so still yet to be finished), but eventually this will comprise of various images of different graphs, and the layout of each level.
Next on the list, I am intending to meet my primary supervisor this Thursday to demo what I have done (and prove its not all an elaborate and well planned lie). After, assuming all goes well and minimal corrections/changes, I will be focusing on a dynamic version of the algorithm similar to Todd Veldhuizen's work on Ubigraph (very inspirational work, also a good indicator of what Im trying to achieve).
Below are some useful links, I hope to update soon with my testing documentation and example outputs. Adiós!
Here is a link to the document mentioned above, a brief description of my implementation. It should be noted this is likely to change over the next few days as I update it to suit my own specification.
Also, I failed to upload this previously, a presentation regarding my Eades implementation and output. Some of this may also be outdated as I have progressed my implementations.
No comments:
Post a Comment