Given that this is my main hobby, I have a plan that should keep me
occupied. Currently my V7 solver is complete with respect to its first of
three passes. If a game doesn’t solve with pass one, it will be either a
true negative, or a false negative. Pass one doesn’t generate any unsolved.
It now operates at 10,000 games per 26 seconds and for the first 1000 games
it gets a median solution length of 46 steps.
I am now working on pass two which I intend to get no false negatives, but a
few unsolvables. Pass three will attempt to solve the remaining
unsolvables. My solver gives each move a type, and scores each move in the
context of its layout. It is driven by rules to prune and prioritize each
move type. It is designed as a 5-state FSM. I have designs in mind for
most of the remaining development. Once I’m done I thought I’d verify the
first 100,000,000 games, then soldier on into the next 100,000,000, and so
on until I’ve gotten through the entire 8,000,000,000+ games of the random
number generator. That will obviously require a lot of compute time.
Before I do that, however, I’ll probably try to get my code working on the
latest Microsoft platform. I think I’ll be at this for another 10 years at
least. :-)
From: dannyjones183
Sent: Tuesday, October 02, 2012 2:10 PM
To: fc-solve-discuss_at_yahoogroups.com
Subject: Re: Status of 8x4 FreeCell Deals
>
> > * After a recent determination of deals #24,795,893 and #53,687,601 as
> > unsolveable, the solveability of the first 100,000,000 deals have been
> > determined.
>
> Would you care to share the list of impossible/unsolveable deals?
I didn't make the run of 100,000,000 deals. Ryan Miller did it a few years
back and provided a list of his intractable deals for me to check. I only
know which of his intractables are unsolveable. The two deals above were the
only hold-outs.
>
> > * For the first 25,000,000 deals, I've been able to solve all but
> > three of the solveable deals in less than 60 moves.
>
> BTW, by "less than 60 moves", do you mean "59 or less" or
> do you mean "60 or less"?
>
59 moves or less. I just completed a recheck of the first 1,000,000 deals to
verify this statement. I must admit that I don't recall doing it for the
remaining deals in the first 25,000,000.
>
> Have you verified that all these solutions of your are minimal? What do
> you want to do further down the road?
>
Only a few of my solutions are minimal. Most only qualify as "fairly short".
However, for a specific deal, I have a solver that can produce a very short
solution, but consumes several minutes in the process.
I don't have any idea where to go from here. That's why I asked.
> I have indirect access to High Performance Computing (HPC) machines with
> 128 GB of RAM (which works reliably) and with 512 GB of RAM (which does
> not
> work very reliably), and can run my solvers (or solvers that are built
> from other supplied sources) there. Would that help you somehow?
>
Not at the moment. I don't have any projects remaining to resolve.
Regards, Danny A. Jones
Received on Tue Oct 02 2012 - 15:50:02 IST