[EM] Cloneproof SSD

[MIKE OSSIPOFF]

Markus wrote:

>you wrote (15 Jan 2001):

> >Aside from criteria,
> > BeatpathWinner's only obvious motivation is that it's an
> > implementation of Cloneproof SSD.
>
>Could you please post an example where these two heuristics
>lead to different winners?

I don't have such an example. That's what I'm asking. Can
Cloneproof SSD give a different result than BeatpathWinner? Is
there any reason why BeatpathWinner is better than Cloneproof SSD,
in terms of results?

>
>If these two heuristics always lead to the same winners then I would
>like to ask: Why should we concentrate on only one of these two
>heuristics?
>
>I always considered it to be advantageous when there is more than one
>heuristic for a given method so that it is possible to explain some
>properties with the one heuristic and other properties with the other
>heuristic. Actually I don't even see why we should differ between
>the Schwartz set heuristic for the Schulze method and the beat path
>heuristic for the Schulze method:

All that's true. It's just that the Schwartz set version is more
publicly proposable. And that gives us incentive to use the Schwartz
version as an official organizational count-rule too, because of
the great value of making a precedent for other uses, leading to
a public proposal. So there's a good reason to use the Schwartz
version because it sets a useful precedent for public proposals.

>Example 1: In so far as the Schwartz set heuristic is an iterative
>heuristic while the beat path heuristic is not an iterative
>heuristic, it is more simple to explain the fact that the
>Schulze method meets monotonicity when one uses the beat path
>heuristic than when one uses the Schwartz set heuristic.

Absolutely.

>Example 2: In so far as the runtime of the Schwartz set
>heuristic is O(NumberOfCandidates^5) while the runtime of
>the beat path heuristic is O(NumberOfCandidates^3), one
>would implement the beat path heuristic even when one used
>the Schwartz set heuristic for the Schulze method to explain
>this method.

Ok, if the number of candidates is great enough to cause significant
delay when using the Schwartz approach. But, again, for the sake of
useful public precedent, even if you use the Beatpath rule, an
organization could do a valuable public service if its procedural rules
mention the Schwartz set rule too, because then people advocating
Cloneproof SSD could point to that organization's procedureal rules and
say "This organization has been using Cloneproof SSD for some time,
and it's worked excellently, and they're quite satisfied with it."

So my main question is: Is there any reason why the beatpath
implementation is better than the Schwartz set implementation, in terms
of results? Is there any difference in results between the 2
implementations?

Mike Ossipoff




>
>Markus Schulze
>
>

_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com