In a bid to keep future posts as short as possible, I am trying to update this more often. As far as can be told, the algorithm for determining disjoint graphs works with what I now to believe O(|V| + |E|) at worst. Too much time has been spent on this so moving on.
I very much enjoy Fruchterman and Reingold's method of modelling particle physics, and thought about how I could utilise some physics myself, such as Entropy as a cooling method and global force acting on every vertex, instead of the annealing method used by Walshaw and Veldhuizen. This obviously needs more research and may prove to be less efficient, then again, it may not, we cant know until we try (or see if anyone else has done it).
Need to do more research on Todd L. Veldhuizen's Dynamic Multilevel Graph Visualisation to see how his algorithm acts on large changes higher in the vertex list (so whilst others are added, what happens if a large number of original vertices dissapears).
No comments:
Post a Comment