Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.python > #88065

Re: Sudoku solver

References (2 earlier) <87fv8sndw1.fsf@elektro.pacujo.net> <CAPTjJmrq4+W7vgVmZZKnOCFtTy5R27vinD1mFUTEyhaUKA+BuQ@mail.gmail.com> <55141449.1030600@davea.name> <CAPTjJmr7p1wYhefJz4RGvhO4F8sTc7tT6BKtGy3m9xrPV+aegg@mail.gmail.com> <55141FB4.4040900@davea.name>
Date 2015-03-27 02:21 +1100
Subject Re: Sudoku solver
From Chris Angelico <rosuav@gmail.com>
Newsgroups comp.lang.python
Message-ID <mailman.207.1427383764.10327.python-list@python.org> (permalink)

Show all headers | View raw

On Fri, Mar 27, 2015 at 2:03 AM, Dave Angel <davea@davea.name> wrote:
> On 03/26/2015 10:41 AM, Chris Angelico wrote:
>  that's already been proven. So, that's why I would avoid guessing.
>>
>>
>> I've written a lot of solvers for various puzzles. Minesweeper,
>> Sudoku, a binary Sudoku-like puzzle that I don't really have a good
>> name for, several others. Every time, I've tried to prove the puzzles
>> solvable by humans, and sometimes that means rejecting ones that could
>> technically be solved by brute force.
>
>
> OK, we're on the same page.  I would use different terminology for some of
> it, but that's okay.

Good, I thought we were in agreement. And yeah, terminology is tricky.

> The purist in me would like to write a solver which (within a few seconds)
> could solve any unique puzzle, and identify puzzles which don't have a
> unique solution.  One reason I never got back to writing one was I also
> wanted a difficulty-ranker, which would identify how hard a human was likely
> to consider the puzzle.

It'd be pretty straight-forward. You simply define a number of
rule-difficulty-categories, something like this:

EASY
* If eight digits are "locked off" from a cell by already existing in
its neighbours, the remaining digit must be the one.
* If eight cells in a {row, column, square} have a given digit locked
off, then the remaining cell must contain that digit.

MEDIUM
* Construct "possibility regions" for digits in squares (ie all the
possible cells that that digit could be in). Any such region which
sits within a single row/column is equivalent to an actual digit in
that row/column for the purposes of exclusions.
* Likewise the converse - if both possible places for the 7 in this
row are in this square, we can depend on the 7 being in one of those,
and not anywhere else in the square.

HARD
* Any complex rule that you feel like coding up. There are plenty of
"Sudoku help" web sites out there that can provide rule definitions.

IMPOSSIBLE
* Brute force. Attempt to put a digit in a cell, then see if you can
solve the puzzle thereafter. If you prove the puzzle to have no
solutions, that digit cannot be in that cell.

Then, in your solver, you use EASY rules until they provide you with
no more material, and only then move on to MEDIUM rules, etc. The
highest difficulty class that you had to use in solving the puzzle is
the puzzle's difficulty class.

This isn't a perfect system, of course, but it's a decent start. It
also deals with the terminology problem: you can declare a puzzle
"solvable, but IMPOSSIBLE class difficulty", which means you have to
guess. Though I would strongly suggest disabling brute-forcing most of
the time, as it'll kill your CPU... especially if you have a puzzle
generation algorithm that looks like this:

Start with a blank grid.
While puzzle not solvable:
    Add random clue digit
For each clue digit:
    Remove clue
    If puzzle not solvable: Reinstate clue

With IMPOSSIBLE deactivated, this is a reasonably straight-forward way
to generate puzzles (and you can deactivate HARD to require a
medium-difficulty puzzle, etc). Otherwise... kerchug, kerchug,
kerchug....

ChrisA

Back to comp.lang.python | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-25 13:39 +0200
  Re: Sudoku solver John Ladasky <john_ladasky@sbcglobal.net> - 2015-03-25 11:44 -0700
    Re: Sudoku solver Ian Kelly <ian.g.kelly@gmail.com> - 2015-03-25 13:05 -0600
    Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-25 21:37 +0200
      Re: Sudoku solver Ian Kelly <ian.g.kelly@gmail.com> - 2015-03-25 14:09 -0600
        Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-25 22:31 +0200
          Re: Sudoku solver Chris Angelico <rosuav@gmail.com> - 2015-03-26 09:40 +1100
          Re: Sudoku solver Ian Kelly <ian.g.kelly@gmail.com> - 2015-03-25 18:07 -0600
            Re: Sudoku solver BartC <bc@freeuk.com> - 2015-03-27 20:04 +0000
          Re: Sudoku solver Abhiram R <abhi.darkness@gmail.com> - 2015-03-26 08:26 +0530
          Re: Sudoku solver Ian Kelly <ian.g.kelly@gmail.com> - 2015-03-25 21:24 -0600
          Re: Sudoku solver Abhiram R <abhi.darkness@gmail.com> - 2015-03-26 08:58 +0530
            Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-26 14:13 +0200
              Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-26 16:15 +0200
          Re: Sudoku solver albert@spenarnc.xs4all.nl (Albert van der Horst) - 2015-04-10 23:10 +0000
      Re: Sudoku solver Sayth <flebber.crue@gmail.com> - 2015-03-27 20:39 -0700
        Re: Sudoku solver BartC <bc@freeuk.com> - 2015-03-28 23:50 +0000
          Re: Sudoku solver Chris Angelico <rosuav@gmail.com> - 2015-03-29 11:12 +1100
            Re: Sudoku solver BartC <bc@freeuk.com> - 2015-03-29 21:59 +0100
              Re: Sudoku solver Mark Lawrence <breamoreboy@yahoo.co.uk> - 2015-03-29 22:19 +0100
                Re: Sudoku solver BartC <bc@freeuk.com> - 2015-03-29 22:59 +0100
          Re: Sudoku solver Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2015-03-29 14:06 +1100
            Re: Sudoku solver Chris Angelico <rosuav@gmail.com> - 2015-03-29 15:10 +1100
              Re: Sudoku solver Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2015-03-29 21:35 +1100
                Re: Sudoku solver Chris Angelico <rosuav@gmail.com> - 2015-03-29 21:54 +1100
                Re: Sudoku solver BartC <bc@freeuk.com> - 2015-03-29 13:01 +0100
                Re: Sudoku solver BartC <bc@freeuk.com> - 2015-03-29 16:23 +0100
            Re: Sudoku solver Christian Gollwitzer <auriocus@gmx.de> - 2015-03-29 09:57 +0200
              Re: Sudoku solver BartC <bc@freeuk.com> - 2015-03-30 11:54 +0100
            Re: Sudoku solver BartC <bc@freeuk.com> - 2015-03-29 12:25 +0100
              Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-29 21:03 +0300
                Re: Sudoku solver Mark Lawrence <breamoreboy@yahoo.co.uk> - 2015-03-29 19:26 +0100
                Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-29 21:33 +0300
                Re: Sudoku solver BartC <bc@freeuk.com> - 2015-03-29 22:11 +0100
                Re: Sudoku solver Ian Kelly <ian.g.kelly@gmail.com> - 2015-03-30 00:50 -0600
                Re: Sudoku solver Christian Gollwitzer <auriocus@gmx.de> - 2015-03-30 09:13 +0200
                Re: Sudoku solver Ian Kelly <ian.g.kelly@gmail.com> - 2015-03-30 01:29 -0600
                Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-30 12:16 +0300
                Re: Sudoku solver Dave Angel <davea@davea.name> - 2015-03-30 04:16 -0400
                Re: Sudoku solver Ian Kelly <ian.g.kelly@gmail.com> - 2015-03-30 02:57 -0600
                Re: Sudoku solver Chris Angelico <rosuav@gmail.com> - 2015-03-30 20:13 +1100
          Re: Sudoku solver Mark Lawrence <breamoreboy@yahoo.co.uk> - 2015-03-29 22:21 +0100
            Re: Sudoku solver BartC <bc@freeuk.com> - 2015-03-29 23:17 +0100
              Re: Sudoku solver Seymore4Head <Seymore4Head@Hotmail.invalid> - 2015-03-29 21:40 -0400
  Re: Sudoku solver Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2015-03-26 09:50 +1100
    Re: Sudoku solver Mark Lawrence <breamoreboy@yahoo.co.uk> - 2015-03-25 23:04 +0000
      Re: Sudoku solver Christian Gollwitzer <auriocus@gmx.de> - 2015-03-27 07:40 +0100
  Re: Sudoku solver "Frank Millman" <frank@chagford.com> - 2015-03-26 10:19 +0200
    Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-26 14:26 +0200
      Re: Sudoku solver Chris Angelico <rosuav@gmail.com> - 2015-03-26 23:37 +1100
      Re: Sudoku solver Ian Kelly <ian.g.kelly@gmail.com> - 2015-03-26 07:15 -0600
        Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-26 16:23 +0200
          Re: Sudoku solver Ian Kelly <ian.g.kelly@gmail.com> - 2015-03-26 09:06 -0600
            Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-26 17:48 +0200
              Re: Sudoku solver Ian Kelly <ian.g.kelly@gmail.com> - 2015-03-26 10:20 -0600
                Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-26 18:47 +0200
      Re: Sudoku solver Dave Angel <davea@davea.name> - 2015-03-26 10:14 -0400
      Re: Sudoku solver Chris Angelico <rosuav@gmail.com> - 2015-03-27 01:41 +1100
      Re: Sudoku solver Dave Angel <davea@davea.name> - 2015-03-26 11:03 -0400
        Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-26 17:42 +0200
          Re: Sudoku solver Pete Forman <petef4+usenet@gmail.com> - 2015-03-26 22:24 +0000
      Re: Sudoku solver Chris Angelico <rosuav@gmail.com> - 2015-03-27 02:21 +1100
      Re: Sudoku solver Dennis Lee Bieber <wlfraed@ix.netcom.com> - 2015-03-26 20:42 -0400
      Re: Sudoku solver Dave Angel <davea@davea.name> - 2015-03-27 09:14 -0400
      Re: Sudoku solver Chris Angelico <rosuav@gmail.com> - 2015-03-28 00:25 +1100
      Re: Sudoku solver "Frank Millman" <frank@chagford.com> - 2015-03-27 15:35 +0200
        Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-27 15:56 +0200
          Re: Sudoku solver Dave Angel <davea@davea.name> - 2015-03-27 10:09 -0400
            Re: Sudoku solver sohcahtoa82@gmail.com - 2015-03-27 11:18 -0700
              Re: Sudoku solver Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2015-03-28 12:46 +1100
            Re: Sudoku solver Larry Hudson <orgnut@yahoo.com> - 2015-03-27 16:48 -0700
          Re: Sudoku solver Chris Angelico <rosuav@gmail.com> - 2015-03-28 01:07 +1100
          Re: Sudoku solver Chris Angelico <rosuav@gmail.com> - 2015-03-28 01:19 +1100
            Re: Sudoku solver Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2015-03-28 14:13 +1300
            Re: Sudoku solver Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2015-03-28 12:40 +1100
              Re: Sudoku solver Ian Kelly <ian.g.kelly@gmail.com> - 2015-03-28 02:24 -0600
          Re: Sudoku solver Mark Lawrence <breamoreboy@yahoo.co.uk> - 2015-03-27 16:03 +0000
          Re: Sudoku solver Virgil Stokes <vs@it.uu.se> - 2015-03-28 19:36 +0100
      Re: Sudoku solver Dave Angel <davea@davea.name> - 2015-03-27 09:48 -0400
      Re: Sudoku solver Dave Angel <davea@davea.name> - 2015-03-27 09:52 -0400
      Re: Sudoku solver Chris Angelico <rosuav@gmail.com> - 2015-03-28 00:58 +1100
  Re: Sudoku solver mr.smittye@gmail.com - 2015-03-29 16:39 -0700
    Re: Sudoku solver Marko Rauhamaa <marko@pacujo.net> - 2015-03-30 12:24 +0300

csiph-web