Hi there,
I was able to write my own FreeCell solver.
It's pretty fast and is able to find good solutions in a very short amount
of time.
My approach uses Beam Search and Zobrist hashing, in around 600 lines of C.
The larger the Beam Search, the better the solutions.
For example, for FreeCell #1, the program finds a solution in 35 moves with
a BS of 10000 (in 2 seconds), and in 32 moves with a BS of 1000000 (240
seconds).
Yuri's solution takes 40 moves.
In 2013, Danny Jones and Yuri Bortnik seem to have tried to optimally solve
the first 1000 games.
Have these results been published somewhere ?
JC
Received on Thu May 28 2015 - 23:55:04 IDT