# Re: Write a Fortran program to solve Sudoku puzzles :-)

*From*: "Michael Metcalf" <michaelmetcalf@xxxxxxxxxxxxxx>*Date*: Fri, 02 Dec 2005 13:28:10 GMT

"AN O'Nymous" <a_n_onymous80@xxxxxxxxxxx> wrote in message

news:1133492643.166008.295960@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

> Interesting, Michael. What technique (e.g. Tabu, simulated annealing,

> partial enumerative, ...?) does your algorithm use to solve it, and

> what decision rules do you use to get the algorithm to give up looking

> for multiple solutions?

>

I apply various logic solving algorithms, for instance 'pairs', and resort

to brute force if the logic rules prove insufficient. Brute force stops as

soon as it finds a second solution (hardly counts as a decision rule - going

on is simply uninteresting). For a human solver, brute force is equivalent

to having to guess (Thread of Ariadne). The line between the two is fuzzy,

depending on how many algorithms you're willing to code and apply (X-wing,

swordfish, jellyfish etc.), just as a human solver may guess rather than

apply such techniques.

I'm planning to write an article for Fortran Forum on this.

Regards,

Mike Metcalf

.

**References**:**Write a Fortran program to solve Sudoku puzzles :-)***From:*AN O'Nymous

**Re: Write a Fortran program to solve Sudoku puzzles :-)***From:*Michael Metcalf

**Re: Write a Fortran program to solve Sudoku puzzles :-)***From:*Michael Metcalf

**Re: Write a Fortran program to solve Sudoku puzzles :-)***From:*AN O'Nymous

- Prev by Date:
**Re: Write a Fortran program to solve Sudoku puzzles :-)** - Next by Date:
**Re: compilation cascades (yet again)** - Previous by thread:
**Re: Write a Fortran program to solve Sudoku puzzles :-)** - Next by thread:
**Re: Write a Fortran program to solve Sudoku puzzles :-)** - Index(es):