[EM] Nondeterminism in Multiwinner Methods

[Raph Frank]

On Sun, Oct 26, 2008 at 4:44 AM, Greg Nisbet <gregory.nisbet at gmail.com> wrote:
> It is unlikely that a nondeterminstic solution would be perfect, of course.
>
> However, I suspect that it can deliver at least some of the benefits
> of group (1) without incurring factorial execution time.
>
> Any thoughts on the matter?

Another option is to allow the candidates to submit winning sets.

Only those winning sets are then considered.

The process would be:

- voters cast their ballots
- ballot info is published*
- Each candidate submits a winning set within (say) 1-2 days
- all candidate sets are considered and the winning set is declared the winner

*This would have to be enough to work out the true winner

This is deterministic (though it depends on both the ballots and the
outcomes submitted by the candidates), but still allows the more open
system to be used.

Each candidate would submit the winning set with the highest score (or
likely Condorcet winner) that has the candidate as a member.

This means that it is very likely that the final outcome is similar to
the true winner.

It also allows unsecured computing power to be used to find a winner.
There could be a "Green Party at Home" effort.  Checking a result is
likely to be much easier than finding the optimal.

Mostly, though everyone will just agree on the winning set, it is only
in edge cases that there is a dispute.