[EM] Alternative algorithms for Yee diagrams.

[Jameson Quinn]

Just eyeballing the pictures, it seems that the radii are very
conservative. If you doubled them, it would be 4 times as fast. I realize
you want to stay on solid ground mathematically, but given how conservative
things look, I think you probably could. Can you build the Hessian matrix
of the tightest decisive vote margin? If so, would taking things to second
order allow a less-conservative first-order approximation? (Just wild
ideas, I'm not sure what I'm saying even makes sense.)


2013/12/11 Leon Smith <leon.p.smith at gmail.com>

> Here's the result of a conversation I had with Claude Heiland-Allen today
> about faster algorithms for generating Yee Diagrams.
>
>
> http://mathr.co.uk/blog/2013-12-11_distance_estimation_for_voting_simulation_visualisation.html
>
> I haven't run any timing tests yet,  so it's not clear if this is faster
> as-is or not.   And there is probably substantial room for further
> improvement... e.g. in selecting starting points more intelligently.
>
> Also, the obvious generalization of this algorithm to ranked methods would
> also slow it down a bit.  But I thought it might be of interest to some
> people here on this list.
>
> Best,
> Leon
>
> ----
> Election-Methods mailing list - see http://electorama.com/em for list info
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20131215/ccde7788/attachment-0003.htm>