- Analysis of Runtime - much of the runtime analysis has been written into the draft chapters of my Thesis, with the general idea that my algorithm is faster than a quadtree and equivalent to a dual tree - with the hopeful benefit of having better quality (equal to that of the quadtree or better). Much of the runtime results for experimentation have also been written up.
- Analysis of Quality - oddly enough the more time consuming task (due to my unnecessary and possibly wasteful time researching speedy methods for counting edge crossings). Thankfully, after some aid from Dr Chris Walshaw and some Googling, I have implemented a far more efficient method of counting edge crossings and reduced the time required by more than half of my previous methods. Unfortunately this still takes a lot of time for huge graphs but better 20minutes than an hour... With this done, I am not awaiting results so I can write them up and hopefully get my first chapter ready for checking.
- Dynamic framework - having designed and implemented a chaco-style method of importing "operations" into the application, I now have a simplistic dynamic graph drawing algorithm ready and working - featuring changes based on the users input (through the operations file) and some adaptability in terms of when to update the layout etc. Still needs a lot of work though, however, the main concept is there - next I will be refining this and implementing a means of measuring the dynamic matching I've been waiting to complete. Also, there are some peculiar behaviours I would like to investigate - so this may inhibit plans. I would also like to implement a similar method of graph drawing to that published by Veldhuizen - so we will see how it goes.
Other various tasks have been completed but I cant remember them, so hopefully nothing important. I will update again when I have more information regarding anything of the above, for now, worky worky work work, lots to do and not enough time to do it - challenge accepted!
No comments:
Post a Comment