[algogeeks] Re: Sudoku Solver

2011-12-06 Thread shady
backtracking On Dec 6, 5:05 pm, Vishnutej mylavarapu.vishnu...@gmail.com wrote: Can anyone tell me how to write a generic code for any board size of sudoku? I mean we can write code for a normal 9x9 sudoku board..What about 16x16,25x25...?? -- You received this message because you are

[algogeeks] Re: Sudoku

2011-10-05 Thread Don
My method does use backtracking, but only when the first two methods fail. When the cell-based and group-based analysis do not produce any results, it picks one cell and tries each possible value in that cell. If that results in a contradiction, it backs out to that point and tries something else.

[algogeeks] Re: Sudoku

2011-10-04 Thread Don
When you say Simulate Sudoku, do you mean solve a given Sudoku problem? Here is an overview of how I did that: I used an array of 81 integers to represent the board. Then I built a 27x9 table of all the groups: 9 rows, 9 columns, and 9 squares. Then I built a 81x3 map which relates each location

Re: [algogeeks] Re: Sudoku

2011-10-04 Thread abhishek sharma
Hi Don, How is your method better than backtracking? i hope you are implementing a sudoku solver.. Let me know. Regds. On Tue, Oct 4, 2011 at 10:42 PM, Don dondod...@gmail.com wrote: When you say Simulate Sudoku, do you mean solve a given Sudoku problem? Here is an overview of how I did

[algogeeks] Re: Sudoku

2011-06-07 Thread Navneet Gupta
A little different but related problem is to determine whether a Sudoku puzzle is easy, medium or hard to solve. Here, solveability is guaranteed. On Tue, Jun 7, 2011 at 2:31 PM, Navneet Gupta navneetn...@gmail.com wrote: What could be a good strategy to check if the given Sudoku puzzle is

[algogeeks] Re: Sudoku verification problem

2011-05-31 Thread Don
A data structure could be a mapping from each of the N*N cells to the three regions it belongs to, and then for each region, a set of N bits indicating which values it already contains. Then the validation simply involves iterating over the cells. For each one, if any of the three regions already

[algogeeks] Re: Sudoku verification problem

2011-05-29 Thread Dumanshu
here in this part if bitmap[bi] matrix[i][j]) 0x1) == 0x0) (((bitmap[bj] matrix[i][j]) 0x1) == 0x0) (((bitmap[bk] matrix[i][j]) 0x1) == 0x0)) { bitmap[bi] |= 1 matrix[i][j];

Re: [algogeeks] Re: Sudoku verification problem

2011-05-29 Thread Vishal Thanki
yes, you are right. bitmap will be filled in the process of solving the grid. in verify routine, if the expression evaluates to false, it mean an element is encountered which is already present in row, col and 3x3 cube. this way you an tell that the solution is wrong. hope that helps. On Sun,

[algogeeks] Re: Sudoku verification problem

2011-05-29 Thread Dumanshu
no... what i mean to say is you have filled the bitmap in the process of solving the grid. The way you are filling the bitmap is - you are taking values from the matrix and placing them in the bitmaps using bi bj bk. now in the verification you are checking the same matrix value against the same

[algogeeks] Re: Sudoku verification problem

2011-05-29 Thread Dumanshu
I think ur verification function is correct but it works only if user is entering the values one by one. as soon as he enters a duplicate value it will show a error. It will work for the case of ur solver as ur code itself is generating the values. But here in this verification problem u r given a

Re: [algogeeks] Re: Sudoku verification problem

2011-05-29 Thread Vishal Thanki
sorry, i made a mistake in explaining. when you are solving the gird, you will have the matrix[][] array partially filled. the bitmap used for solving the grid is different than the one being used for verifying. in case of verifying, you will have a completely blank bitmap, with not bits set. you

Re: [algogeeks] Re: Sudoku verification problem

2011-05-29 Thread Vishal Thanki
what i understood from your problem is, when the configuration is given by user, it can be directly stored into matrix[][] array. that will solve your problem in o(n^2), where n=9. On Sun, May 29, 2011 at 5:50 PM, Dumanshu duman...@gmail.com wrote: I think ur verification function is correct but

[algogeeks] Re: Sudoku verification problem

2011-05-29 Thread Dumanshu
oh yeah... now it works. thanx a lot! On May 29, 5:20 pm, Vishal Thanki vishaltha...@gmail.com wrote: sorry, i made a mistake in explaining. when you are solving the gird, you will have the matrix[][] array partially filled. the bitmap used for solving the grid is different than the one being

[algogeeks] Re: Sudoku Algorithm

2008-01-01 Thread adak
No help from here on the difficulty calculation for Sudoku solving. My solver works like most of them, I'd guess. From the starting position: 1) Add any must have numbers, if any, to the set of known square values. 2) Delete any can't be numbers from the list of possible numbers for each