Re: Bidirectional Binary Self-Joins

On Apr 1, 7:49 am, paul c <toledobythe..._at_oohay.ac> wrote:
>
> > What's surprising is that there are now getting to be pretty decent
> > approximate TSP solutions...
>
> I doubt that helps much. All those "subproblems" add up. Given their
> supposed requirement, they had two choices 1) at quotation time, with an
> origin and destination and some arbitrary limit on the number of legs,
> say 4 or 5, evaluate the freight costs for each leg to any of the 5,999
> other airports which would have made quite an explosive activity for any
> dbms while the customer waited on the 'phone, or 2) populate at least
> one thirty-six million row table based on the same calculation for every
> possible origin and destination every night for use the next day.

I think his point was that there are cheap *approximations* of TSP. I have vaguely heard that these work pretty darn well.

However 2) was the first thing I thought of. And perhaps surprisingly 2) doesn't seem very hard to me. That's the solution I'd investigate. How much computing power does that need? How much does a plane cost?</rhetorical>

Marshall Received on Sun Apr 01 2007 - 17:14:28 CEST