[EM] A computationally feasible method

[Raph Frank]

On Mon, Sep 1, 2008 at 10:58 PM, Kristofer Munsterhjelm
<km-elmet at broadpark.no> wrote:
> In multiwinner election situations (like CPO-STV), the randomness might make
> the losers complain that they lost because the assembly that the
> optimization algorithm stumbled on didn't include them, not because the
> optimal assembly didn't include them. The "everyone may propose a solution"
> approach would to some extent limit these complaints - those who won could
> say "well, if you're on the optimal assembly, why didn't you calculate it
> and submit it as proof?" - but not eliminate it altogether.

If there is a condorcet winner, then it should be OK.  Each
party/candidate could submit a result.  They could be quickly checked
against each other and the condorcet winner of those submitted
determined.

However, if there is a circular tie, would that still work?  I guess
it depends if the method meets independence of irrelevant
alternatives, so that parties don't think "if I had submitted this
other result, then we would have won", unless that result is actually
superior to the previous winner.  In which case, they should have paid
more for supercomputer resources.