EUROPEAN ACADEMIC RESEARCH, VOL.

I, ISSUE 6/ SEPEMBER 2013

ISSN 2286-4822, www.euacademic.org
IMPACT FACTOR: 0.485 (GIF)

A Novel Algorithmic approach for solving Sudoku

puzzle in Guessed Free Manner

ARNAB K. MAJI
Department of IT, North Eastern Hill University
Shillong, Meghalaya
India
SUDIPTA ROY
Department of IT, Assam University
Silchar, Assam
India
RAJAT K. PAL
Department of CSE, University of Calcutta
Kolkata, West Bengal
India

Abstract:
"Sudoku" is the Japanese abbreviation of a longer phrase,
"Suuji wa dokushin ni kagiru", meaning "the digits must remain
single". It is a challenging numeric puzzle that trains our logical mind.
Solving a Sudoku puzzle requires no math, not even arithmetic. Even
so, the game poses a number of intriguing mathematical problems. The
problem of solving a given Sudoku puzzle finds numerous applications
in practice. All the existing Sudoku solving techniques are primarily
guess based heuristic or computation intensive soft computing
methodology. In case of solving 9x9 Sudoku Puzzle, in each of these
algorithms, we have to separately go through 81 cells and perform
backtracking for the individual cells. In this paper, an attempt has
been made to develop an algorithm which is minigrid based, i.e., we
have to individually go through nine minigrids (instead of 81 cells)
and perform backtracking only on them, which is less time consuming.
Moreover, no guessing is involved in the whole computation.

1126
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

Key words: Sudoku, Cell, Grid, Difficulty Level, Algorithm,

Backtracking, Permutation, Elimination.

1. Introduction:

A Sudoku is usually a 9×9 grid based puzzle problem

which is subdivided into nine 3×3 minigrids, wherein some
clues are given and the objective is to fill it up for the remaining
blank positions. Furthermore, the objective of this problem is to
compute a solution where the numbers 1 through 9 will occur
exactly once in each row, exactly once in each column, and
exactly once in each minigrid independently obeying the given
clues. One such problem instance is shown in Figure 1(a) and
its solution is shown in Figure 1(b).

6 9 8 3 7 6 2 5 1 4 9 8
2 5 1 2 5 9 4 8 6 7 1 3
7 5 8 1 4 3 7 9 6 2 5
5 6 3 4 7 9 5 6 2 3 1 8 4
4 8 2 9 1 7 5 3 6
1 8 5 2 9 1 6 3 8 4 5 2 7 9
9 6 9 3 7 1 6 4 8 5 2
2 4 7 6 2 1 5 9 8 3 4 7
5 4 2 9 5 4 8 7 3 2 9 6 1

Figure 1: (a) An instance of the Sudoku problem. (b) A solution of the

Sudoku instance shown in 1(a).

Besides the standard 9×9 grid, variants of Sudoku

puzzles include the following:
• 4×4 grid with 2×2 minigrids,
• 5×5 grid with pentomino regions published under the name
Logi-5. A pentomino is composed of five congruent squares,
connected orthogonally. Pentomino is seen in playing the game
Tetris,
• 6×6 grid with 2×3 regions,
EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013
1127
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

• 7×7 grid with six heptomino regions and a disjoint region,

• 16×16 grid (super Sudoku),
• 25×25 grid (Sudoku, the Giant),
A complete Sudoku solution grid may be arrived at in
more than one way, as we can start from any given clues that
are distributed over the minigrids of a given incomplete grid.
Nobody has yet succeeded in determining how many different
starting grids there are. Moreover, a Sudoku starting grid is
really only interesting to a mathematician if it is minimal, i.e.,
if removing a single number means that the solution is no
longer unique. No one has figured out the number of possible
minimal grids, which amounts to the ultimate count of distinct
Sudoku puzzles. It is a challenge that is sure to be taken up in
the near future.
The Sudoku problem is important as it finds numerous
applications in a variety of research domains with some sort of
resemblance. Applications of solving a Sudoku instance are
found in the fields of Steganography (Hong et al. 2008), Secret
image sharing with necessary reversibility (Chang et al. 2010),
Digital watermarking (Naini et al. 2010), Image authentication
(Wu and Ren 2009), Image Encryption (Wu et al. 2011),
Enhancement of genome sequence in DNA Sudoku (Enrich et
al. 2009), Track maintenance through cooperating agents
(Thaens, 2008) and so and so forth.
By the way, all the earlier existing Sudoku solvers that
are available in literature (and Internet) are entirely guess
based and hence extremely time consuming (Jussien 2007). In
addition, each of these existing solvers solves an instance of the
problem considering the clues one-by-one for each of the blank
locations. Often guessing may not be guided by selecting a
desired path of computing a solution and hence exhaustive
redundant computations are involved over there. On the other
hand, the solver developed in this paper is a minigrid based
guessed free Sudoku solver which is a more deterministic
algorithmic approach in the sense that redundancy is

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1128
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

drastically reduced in this process of computations involved and

that it always guarantees a solution if it exists in a reasonable
amount of time.

2. Literature Review:

An n2×n2 Sudoku grid (consisting of n×n blocks) is an

NP-complete problem (Yato and Seta 2003). So, it is unlikely to
develop a polynomial time algorithm to solve this problem.
There are quite a few logic techniques people use to solve this
problem. Some are basic simple logic, some are more advanced.
Depending on the difficulty of the puzzle, a mixture of
techniques may be needed in order to solve a puzzle.
Usually Sudoku instances are available in literature
based on their classification of different levels of difficulty like
easy, moderate, diabolical, etc. These classifications are based
on the algorithmic approaches developed and executed in
solving different Sudoku instances. Sometimes it is told that
the instances are easier or harder based on the number of clues
given along with their relative locations in a given instance but
there is no proof to support such claims. Incidentally, the
Sudoku solver developed herein does not differentiate the
instances in terms of any level of difficulty, and each of the
instances is equally easy or hard to solve using our approach, if
a reasonable number of clues are given. Table 1 shows a
comparison chart of the number of clues for different difficulty
levels (Lee 2006).
However, position of each of the empty cells also affects
the level of difficulty. If two puzzles have the same number of
clues at the beginning of a Sudoku game, the puzzle with the
givens (or clues) in clusters is graded in higher level than that
with the givens scattered over the space. Based on the row and
column constraints, the lower bound on the number of clues are
regulated in each row and column for each difficulty level (Lee
2006). as shown in Table 2.

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1129
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

Table 1: Number of clues given in a Sudoku puzzle in defining the

level of difficulty of a Sudoku instance.
Difficulty level Number of clues
1 (Extremely Easy) More than 46
2 (Easy) 36-46
3 (Medium) 32-35
4 (Difficult) 28-31
5 (Evil) 17-27

Table 2: The lower bound on the number of clues given in each row
and column of a Sudoku instance for each corresponding level of
difficulty.
Difficulty level Lower bound on the number of
clues in each row and column
1 (Extremely Easy) 05
2 (Easy) 04
3 (Medium) 03
4 (Difficult) 02
5 (Evil) 00

It has already been told that our approach developed in

this paper does not differentiate the instances, rather our
approach computes a solution if it exists without guessing a
possible value in a blank location and minigrid based
irredundant deterministic computations are involved over
there.
The basic technique that has been adopted for solving
Sudoku puzzles is backtracking (Jussien 2007). It works as
follows. The program places number 1 in the first empty cell. If
the choice is compatible with the existing clues, it continues to
the second empty cell, where it places a 1 (in some other row,
column, and minigrid). When it encounters a conflict (which can
happen very quickly), it erases the 1 just placed and inserts 2
or, if that is invalid, 3 or the next legal number. After placing
the first legal number possible, it moves to the next cell and

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1130
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

starts again with a 1. If the number that has to be changed is a

9 (which cannot be raised by one in a standard Sudoku grid),
the program backtracks and increases the number in the
previous cell (the next-to-last number placed) by one. Then it
moves forward until it hits a conflict.
In this way, the program may sometimes backtrack
several times before advancing. It is guaranteed to find a
solution if there is one, simply because it eventually tries every
possible number in every possible location.

Figure 3. (a) An instance of a Sudoku puzzle. (b) Potential values in

each blank cell are inserted based on the given clues of the Sudoku
instance in Figure 3(a); here green digits are naked singles. (c) The
concept of naked singles is preferably used to reduce the domain of
probable candidate values in each blank cell, and the process is
successive in nature to find out consequent naked singles, as much as
possible. As for example, the naked single for cell [9,8] is 2, as 4 and 8
have already been recognized as naked singles along row 9 and
column 8; then 8 is a naked single for cell [7,8], as 2 and 4 are already
identified naked singles along column 8, and so on.

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1131
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

2 6 8 5

4 5 7 2 3

7 5 1

5 4 6 8 2

6 1 7

9 8 7 6 5

7 5 6 9

(a)
1 3 134
9
1 3 2 6 8 79
1 4 5 1 4

4 5 7 1 7
8 91 6
1 9 1 9
2
1 3 1 8 6 8 1 3 4 1 3 4 1 3 4 1 2 2 4 7 18 4 96
6 5 7 9 7 9 5 7 9 8 9 8 9
8 8
8 9
9 4 6
2 8 2 3 2 3 3 4
879 5 1 3 6

7 8 9 41 97 4 9 8 9
2
5 3 4 6 1 7 8 9
27 39
2 8 6 1 4
2 3 37 49
3 4 4 9 7
5
4 9
18 29 1 2 5 6 1 2 3 1 2 1 3 1 2 2 4 1 3
4 8
5 9 8 9
8 9 4 7 9 3 4
6 8 1 2 1 4 3 4 8 4 8
1 2
8 9 498 3 8 4 7 7 9 7 6 5
31 2 1 2 39 4 4 2 4 1 3
7 3 8
5 6 9
3 8 3 4 8 4 8
(b)
1 3 1 3 4
9 1 2 6 8 7 9
1 4 5 1 4

91 6
4 5 7 1 9 1 9 2 1 7 8
1 3 1 3 4 1 3 4 1 3 4 2 4 138 49
1 8 6 8 1 2
6 5 7 9 7 9 5 7 9 7 8 6 8
9 5
93 6
8 79 2 8 8 9 2 3 2 3 3 4 45 6 1
46 81 97 2
5 3 4 4 97
1 9
8 9
2 3 3 4
2 8
6 1 4
2 3
5
3 4 4 7
91 2 1 2 5 6 1 2 3 41 92 1 3
8 94 1 2 2 4 1 3
8 9 4 8 5 9
4 7 9 3
8 3 4 8 4 8
1 2 9 3 68 1 2 1 4
7 9 7 6 5
4 7
(c) 9
4 1 3
1 2
7 8 81 92 5 6 9 2 4

8
EUROPEAN ACADEMIC 3 4I, ISSUE 6/ SEPTEMBER 20138
RESEARCH, VOL.
4 8
1132
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

Some other techniques include elimination based

approach (Lee 2006) and soft computing based approach
(Jussien 2007). Let us now focus to review the elimination
based approach. In this approach, based on the given clues a
list of possible values for every blank cell is first obtained. Then
using the following different methods such as naked single,
hidden single, lone ranger, locked candidate, twin, triplet, quad,
X-wing, XY-wing, swordfish, coloring, we eliminate the multiple
possibilities of each and every blank cell, satisfying the
constraints that each row, column, and minigrid should have
the numbers 1 through 9 exactly once. An instance of a Sudoku
puzzle and its possible values of each blank cell are shown in
Figures 3(a) and 3(b), respectively.

B. Naked single
If there is only one possible value existing in a blank
cell, then that value is known as a naked single (Lee 2006).
After assigning the probable values for each blank cell, as
shown in Figure 3(b), we obtain the naked singles 3, 9, and 3 at
locations [5,2], [5,8], and [8,3], respectively. So, we can directly
assign these values to these cells. Then we eliminate these
digits (or naked singles) from each of the corresponding row,
column, and minigrid. Hence, after elimination of these
numbers, as stated above, we obtain a modified (reduced) status
of each blank cell as shown in Figure 3(c), wherein several
other naked singles could be found (and this process is
recursive until no naked singles are found).

C. Hidden single
Sometimes there are blank cells that do, in fact, have
only one possible value based on the situation, but a simple
elimination of candidate in that cell‟s row, column and minigrid
does not make it obvious. This kind of possible value is known
as a hidden single (Lee 2006). Suppose, if we re-examine the
possible values in each cell of Figure 3(b), hidden single can

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1133
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

easily be found in cell [7,2] whose value must be 4 as in

minigrid numbered 7, 4 is not there as probable values in other
cells. Similarly, for cell [4,9], the hidden single is 6 (as in other
cells of the same minigrid 6 is not present as probable values).
Most of the puzzles ranked as easy, extremely easy and
medium can simply be solved using these two techniques of
singles.

D. Lone ranger
Lone ranger is a term that is used to refer to a number
that is one of multiple possible values for a blank cell that
appears only once in a row, or column, or minigrid (Lee 2006).
To see what this means in practice, consider a row of a Sudoku
puzzle with all its possibilities for each of the cells (red digits
are either givens or already achieved), as shown in Figure 4. In
this row, six cells (with red digits) have already been filled in,
leaving three unsolved cells (second, eighth, and ninth) with
their probable values written in them.

3 6
2 7 4 1 5 9 8 6 7 6 7

Figure 4. An example row of a Sudoku puzzle with a lone ranger 3 in

the second cell.
Notice that the second cell is the only cell that contains
the possible value 3. Since none of the remaining cells in this
row can possibly contains 3, this cell can now be confirmed with
the number 3. In this case, this 3 is known as a lone ranger.

E. Locked candidate
Sometimes it can be observed that a minigrid where the
only possible position for a number is in one row (or column)
within that block, although the position is not fixed for the
number. That number is known as a locked candidate (Lee
2006). Since the minigrid must contain the number in a row (or

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1134
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

column) we can eliminate that number not as a probable

candidate along the same row (or column) in other minigrids.
Consider the Sudoku puzzle along with its probable
assignments for each blank cell, as shown in Figure 5. It can
readily be found that minigrid numbered 6 should have 3 in the
last row. So we can simply eliminate number 3 from cell [6,5] of
minigrid numbered 5. Similarly, minigrid numbered 8 should
have 3 in its first column. So, 3 can be eliminated as a possible
candidate from cell [4,4].

2 4 5 8 7 9 3 4 5 1 6
4 6 4 6
9 9 1 8 2 5 7 4 9 3
3 5
9
5 3 9 1 4 6 8 5 9 2
1 4 4 5
7
3 4 3 6 9 2 8 1 4
5 6 5
6
4 5
3 2 4 5 1 8 9 6 7
145 4 6 3 4 1 4
6 9 8 9
2 3 6 7 5
3 4

5
3 4 2 5 6 8 1 3 7 9
4
4 6 3 4
7 9 6 9
3 9 5 2 1 3 4 8

8 1 3 9 3 9 7 4 6 2 5

Figure 5. A Sudoku puzzle with probable locked candidates in the

last row of minigrid 6 (and here the locked candidates are 3 and 5 in
cells [6,7] and [6,8]), in the first column of minigrid 8 (and here the
locked candidates are 9 and 3 in cells [8,4] and [9,4]), and so on.

F. Twin
If two same possible values are present for two blank
cells in a row (or column) of a Sudoku puzzle, they are referred
to as twin (Lee 2006). Consider the partially solved Sudoku
puzzle as shown in Figure 6(a). Observe the two cells [2,5] and
[2,6]. They both contain the values 2 and 3 (means either 2 or
EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013
1135
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

3). So, if cell [2,5] takes value 2, then cell [2,6] must contain 3,
or vice versa. This type of situation is an example of twin.
Once a twin is identified, these values can be eliminated
by striking through from the same row, column, and minigrid
as shown in Figure 6(b), as the values cannot be probable
candidates in other blank cells along the same row (or column)
and in the same minigrid.
4 7 2 3 2 3 3 4 3 4
5 8
6 8 9 8 9
1 3 4 5 5
2 4 2 4
5
8 5 7 2 3 2 3 51
6 9

9 3 1 6 4 5 2 8 7
2 3
1 2 4
7 5 2 3 7 8 9 4

(a)

4 7 2 3 2 3 3 4
5 8
6 8 9 8 9
1 3 4
5
3 4
5

2 4 5
2 4
5
8 5
7 2 3 2 3 1 6 9

9 3 1 6 4 5 2 8 7

2 3
1 2 4 7 5 2 3 7 8 9 4

(b)
Figure 6. (a) A partial Sudoku instance with presence of twin 2 and 3
in cells [2,5] and [2,6]. (b) Elimination of probable values (that are 2
and 3) based on the twin from the second row (2 is deleted from cells
[2,1] and [2,3]) and from the same minigrid (2 and 3 are deleted from
cells [1,4] and [1,5]).

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1136
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

4 5
7 6
9 8 1 2 3
4 5 4 5
6 6

(a)

4 5 4 5
7 6 9 8 1 2 3 6
4 5

(b)

4 5
7 4 5 9 8 1 2 3 4 6
6

(c)
Figure 7. Example rows of Sudoku puzzles with different varieties of
triplet. (a) A triplet of Variety# 1 with same three possible values
present in three cells. (b) A triplet of Variety# 2 with same three
possible values present in two cells and the other cell containing any
two of them. (c) A triplet of Variety# 3 with three possible values
present in one cell and the two other cells containing two different
subsets of two possible values of the earlier three values.

G. Triplet
If three cells in a row (or column) are marked with a set
of same three possible values, they are referred to as triplet
(Lee 2006). Like twins, triplets are also useful for eliminating
some other possible values for other blank cells. Triplet has
several variations like the following.
Variety# 1: Three cells with same three possible values,
as shown in Figure 7(a).
Variety# 2: Two cells with same three possible values
and the other cell containing any two of the possible
values, as shown in Figure 7(b).
Variety# 3: One cell with three possible values and the
two other cells containing two different subsets of two
possible values of the former three values, as shown in
Figure 7(c).

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1137
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

Once a triplet is found, we can eliminate all the values of

the triplet that are there as possible candidates in other blank
cells along the same row (or column) and in the same minigrid.

H. Quad
Analogous to triplet, a quad consists of a set of four
possible values and these values are present in some form in
four blank cells in a row (or column) of the Sudoku instance
(Lee 2006). That is, if the values only exist in four (blank) cells
in a row (or column), while each cell contains at least two of the
four values, then other values (or numbers except the specified
four values) can be eliminated from each of the assumed cells
(forming the quad). Figure 8 shows a row of a Sudoku puzzle
where the quad comprising the digits 1, 2, 4, 7 formed by the
cells in column four, six, seven, and eight. So other possible
values can straightway be eliminated from these cells, as shown
by striking through the inapplicable digits in the figure.
1 2
1 2 1 2 1 4
3 5 8
4 7
1 8
4 8 7 8
7 8 6
9 9

Figure 8. An example row of a Sudoku puzzle with quad comprising

digits 1, 2, 4, and 7 present in columns four, six, seven, and eight. To
support the digits present in the quad in the stated cells, other
probable values (like 8 and 9 in columns six, seven, and eight) are
eliminated from these cells of the quad, as these values (that are 8
and 9) cannot be probable digits for the specified cells.

An extended version of the above algorithm(Jussien

2007) defines a set of terms like XWing, Swordfish, Hidden
subset, etc, but ultimately it is also a trial based algorithmic
way-out which is a guess based, cell based Sudoku solver
(Jussien 2007). The XWing technique can be applied when there
are two rows or columns for which a given value is possible to
assign only to two blank cells. If those four cells are in only two
orthogonal rows or columns, then all other cells in those regions

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1138
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

will never get assigned to this value. The XWing technique can
easily be generalized but swordfish is a technique that further
makes the possibility of assigning a digit to a blank cell more
specific. On the other hand, hidden subset is a kind of technique
that is very similar to twin or triplet or quad that have already
been explained above.
On the other hand, all the soft computing based Sudoku
solvers either use Genetic algorithm (Mantere and Koljonen
2007) or Bee colony (Pacurib et al. 2009), which is exhaustive
and extremely time-consuming. Needless to mention that these
techniques use their own set of operators to execute the
respective algorithms. Genetic algorithms belong to the larger
class of evolutionary algorithms that generate solutions to
optimization problems using techniques inspired by natural
evolution, such as selection, crossover, mutation, and
inheritance.
The Simulated annealing based Sudoku solver is a
probabilistic Sudoku solver. The general design is capable of
solving a Sudoku instance of order up to fifteen [7]. It has been
claimed that the solver has solved in actual hardware Sudoku
puzzles of up to order 12 within the competition-imposed time
limits.

3. Proposed Algorithm for solving Sudoku puzzle in a

guessed free manner:

All the previous algorithms discussed in the literature

survey are cell based and some amount of guessing is always
involved in all the technique. In this paper an attempt has been
made to develop an algorithm which is minigrid based, i.e., we
have to individually go through nine minigrids (instead of 81
cells) and perform backtracking only on them, which is less
time consuming. Moreover, no guessing is involved and no
redundant computation is performed during the whole
computation.

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1139
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

Our proposed algorithm considers each of the minigrids

that may be numbered as 1 through 9 as shown in Figure 9.
Each minigrid may or may not have some clues as numbers
that are given. We first consider a minigrid that contains a
maximum number of clues, and if there are two or more such
minigrids, we consider the one with the least minigrid number.
Needless to mention that each of the cells in a minigrid,
either containing a clue or a blank cell, is somehow
differentiated from each of the cells of another minigrid as the
position of a cell in a Sudoku instance could be specified by its
row number and column number, which is unique. So, a cell [i,
j] of minigrid k may either contain a number l as a given clue or
a blank location that is to be filled in by inserting a number m,
where 1 ≤ i, j, k, l, m ≤ 9.

1 2 3
4 5 6
7 8 9
Figure 9: The structure of a 99 Sudoku puzzle (problem) with its
nine minigrids of size 33 each as numbered 1 through 9. Minigrid
number 1 consists of the cell locations [1, 1], [1, 2], [1, 3], [2, 1], [2, 2],
[2, 3], [3, 1], [3, 2], and [3, 3], minigrid number 2 consists of the cell
locations [1, 4], [1, 5], [1, 6], [2, 4], [2, 5], [2, 6], [3, 4], [3, 5], and [3, 6],
and so on.

Now to start with a minigrid as stated above, we find

that the minigrid 3 contains a maximum number of clues, i.e.,
4, among all the minigrids, and each of the minigrids 1 and 2
contains less number of clues than that of minigrid 3 (see
Figure 1(a)). For example, for the Sudoku instance as shown in
Figure 1(a), each of the minigrids 3, 5, and 7 contains four clues

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1140
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

each; hence, at the beginning, we consider minigrid 3 for

computing all its valid permutations of the missing numbers for
its blank locations (as 3 is the minimum minigrid number).
Besides, for a given Sudoku instance, we know all the
clues given as well as the clue positions among the cells of a
minigrid and subsequently the blank cells are also known to us.
For example, the given clues in minigrid 3 of Figure 1(a) are 9
at location [1, 8], 8 at location [1, 9], 1 at location [2, 8], and 5 at
location [3, 9]. Here we denote a cell location of a Sudoku
instance by [row number, column number], where each of row
number and column number varies from 1 to 9. Hence the
blank locations are [1, 7], [2, 7], [2, 9], [3, 7], and [3, 8], and the
missing digits are 2, 3, 4, 6, and 7.
In this algorithm, we compute all possible permutations
of these missing digits in minigrid 3, where the first
permutation may be 23467 (the minimum number) and the last
permutation may be 76432 (the maximum number using the
missing digits). Here as the number of blank locations is five,
the total number of permutations is 5!, which is equal to 120.
Now the algorithm considers each of these permutations one
after another and identifies only the valid set of permutations
based on the given clues available in rows and columns in other
minigrids (that are minigrids 1, 2, 6, and 9). As for example, if
we consider the first permutation 23467 and place the missing
digits, respectively, in order in locations [1, 7], [2, 7], [2, 9], [3,
7], and [3, 8], which are arranged in ascending order, we find
that this permutation is not a valid permutation. This is
because the location [6, 7] already contains 2 as a clue of
minigrid 6, and we cannot place 2 at [1, 7] as the permutation
suggests. Also the location [3, 5] contains 7 as a clue of minigrid
2, and we cannot place 7 at [3, 8] as it is supposed to place.
Similarly, we may find that the last permutation 76432
is also not a valid permutation as location [4, 9] already
contains 4 as a clue of minigrid 6, and we cannot place 4 at [2,
9] as the permutation suggests. But we may observe that 74362

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1141
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

is a valid permutation as we may safely place 7 at [1, 7], 4 at [2,

7], 3 at [2, 9], 6 at [3, 7], and 2 at [3, 8] based on the other clues
in the corresponding rows and columns of other minigrids (that
are minigrids 1, 2, 6, and 9).
This is how we may compute all valid permutations of
minigrid 3, and proceed for a next minigrid that belongs to
among the row and column minigrids of minigrid 3 which
contains a maximum number of clues but the minigrid number
is minimum. Among all the valid permutations (for their
respective blank locations) of minigrid 3, at least one
permutation must last at the end of computation of valid
permutations of each of the remaining minigrids if the solution
of the given Sudoku instance is unique. To find out the next
minigrid to be considered, we go through the row and column
minigrids of minigrid 3 in the Sudoku instance of Figure 1(a)
(that are minigrids 1, 2, 6, and 9), and among these minigrids
we find that the minigrid 1 contains a maximum number of
clues, i.e., 3 (which is equally true for each of the minigrids 6
and 9), and its minigrid number is the minimum.
So, now we consider minigrid 1, and as done before for
minigrid 3, we find the given clues and the missing digits
therein along with their locations. Here we do exactly the same
as we did earlier in computing all permutations of the missing
digits in minigrid 3. At the time of identifying all valid
permutations of minigrid 1, we consider one valid permutation
(at their respective blank locations) of minigrid 3 in addition to
all given clues of the instance under consideration. If we get at
least one valid permutation for minigrid 1 (obeying an assumed
valid permutation of minigrid 3), we consider it for some
subsequent computation of permutations of another minigrid;
otherwise, we consider a second valid permutation of minigrid
3, and based on that we compute another set of valid
permutations for minigrid 1, and so on.
Now it is straightforward to declare that here the
minigrid that is to be considered is one among the minigrids 2,

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1142
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

4, 6, 7, and 9 as the row and column minigrids of minigrids 3

and 1 (for which we have already computed valid
permutation(s) one after another); note that neither of
minigrids 5 and 8 is a row or column minigrid of minigrids 3
and 1. Hence following the instance in Figure 1(a), we consider
minigrid 7 for computing all its valid permutations allowing for
one valid permutation of minigrid 3 and then one subsequent
valid permutation of minigrid 1, in addition to all given clues of
the instance under consideration, as each of the minigrids 2, 4,
6, and 9 contains less number of clues than that of minigrid 7.
Here in computing all valid permutations of minigrid 7, we may
not consider an assumed valid permutation of minigrid 3, as
this minigrid is neither in a row nor in a column of minigrid 7,
but we have to consider a valid permutation of minigrid 1 and
all given clues in the Sudoku instance (primarily the clues
given in minigrids 4, 8, and 9).
This process is continued till a valid permutation of a
minigrid (or a set of valid permutations of a group of minigrids)
is propagated to compute a valid permutation of a subsequent
minigrid, and eventually a valid permutation of the last
minigrid (i.e., the ninth minigrid; not necessarily minigrid
number 9) is computed, which altogether generate a desired
solution of the given Sudoku instance. It may so happen that up
to t minigrids t valid permutations that we consider in a series
match each other towards a valid combination of the given
Sudoku instance but there is no valid permutation for the
(t+1)th minigrid obeying the earlier assumed t valid
permutations, where 1  t  9.Then we consider a second valid
permutation of the tth minigrid, and after that we try to
compute a valid permutation for the (t+1)th minigrid, if one
exists. If for none of the valid permutations of the tth minigrid a
valid permutation for the (t+1)th minigrid is obtained, we
consider a second valid permutation for the (t−1)th minigrid that
leads to compute a new set of valid permutations for the tth
minigrid, and so on.

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1143
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

We claim that we must acquire at least one valid permutation

for each of the minigrids one after another, obeying at least one
valid permutation computed for each of the minigrids
considered earlier in the process of assuming the minigrids in
succession; we claim this result in the form of the following
theorem if at least one solution of the given Sudoku puzzle
exists.
Theorem 1: There is at least one valid permutation for
the missing digits for their respective blank locations in each of
the minigrids such that the combination of all such (nine) valid
permutations for all the (nine) minigrids produces a desired
solution, if there exists a solution of a given Sudoku instance.
Proof: The proof of the theorem is straightforward
following the steps of the inherent development of the
algorithm as stated above, if a feasible solution of the given
Sudoku instance is there. We may start with one valid
permutation for some earlier assumed minigrid that may not be
a valid partial solution in combination for the whole Sudoku
instance; then we must reach to a point of computing a valid
permutation of some subsequent minigrid when no such
permutation is obtained for that minigrid. In that case we are
supposed to return back to the former minigrid we had to
consider a next valid permutation, if any, for the same (i.e., for
the previous minigrid) and move to the current minigrid for
computing its valid permutations accordingly. Hence it is clear
that if one valid permutation for some earlier assumed minigrid
is not a valid partial solution in combination for the whole
Sudoku instance, then we must have to return back to that
prior minigrid to consider a new valid permutation of the same
to continue the process again in computing all valid
permutations of its subsequent minigrid, and so on. In this way,
a set of individual valid permutations is to be differentiated so
that in combination of all of them a desired solution of the given
Sudoku instance is computed, if one such solution exists.

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1144
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

To see the algorithm at a glance, let us write it in the form as

follows:
Algorithm: A Guessed Free Sudoku Solver – Version 1
Input: A Sudoku instance, P of size 9×9.
Output: A solution, S of the given Sudoku instance, P.
Step 1: Compute the number of clues, digit(s) given as clue,
and the missing digits in each of the minigrids of P.
Step 2: Compute SM, a sequence of minigrids that contains all
the minigrids in succession, wherein M SM is the minigrid
(and the first member in SM) with a maximum number of clues
and whose minigrid number is minimum. In SM, a member N is
a minigrid which is either in the row or in the column of any of
its earlier members in SM including M that contains a
maximum number of clues and whose minigrid number is
minimum, where 1  N ≤ 9.
Step 3: Compute all valid permutations for the missing digits
in M, and store them.
Step 4: For all the remaining minigrids in succession in SM do
the following:
Step 4.1: Consider a next minigrid, N SM, and compute all its
valid permutations for the missing digits in N assuming a valid
permutation for each of the earlier minigrids up to M, and store
them.
Step 4.2: If one valid permutation for N is obtained, then
consider a next minigrid of N in SM, if any, and compute all its
valid permutations for the missing digits in this minigrid
assuming a valid permutation for each of the earlier minigrids
up to M, and store them.
Else consider a next valid permutation, if any, of the
immediately previous minigrid of N, and compute all its valid
permutations for the missing digits in N assuming a valid
permutation for each of the earlier minigrids up to M, and store
them.
Step 5: If all the valid permutations of the immediate successor
minigrid of M are exhausted to obtain a valid combination for

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1145
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

all the nine minigrids in SM, then consider a next valid

permutation of M and go to Step 4. The process is continued
until a valid combination for all the nine minigrids in SM is
obtained as a desired solution S for P; otherwise, the algorithm
declares that there is no valid solution for the given instance P.
Now it is straightforward to compute SM for a given Sudoku
instance P. As for example, consider the Sudoku instance given
in Figure 1(a). According to this instance the sequence SM of
minigrids is 3, 1, 7, 6, 5, 9, 4, 8, 2 as it has been described and
performed in Step 2 of the first version of the algorithm above.
Computation of all valid permutations for the missing
digits in a minigrid is an important task of the present
algorithm. At the time of computing only all valid permutations
for the missing digits, we follow a tree data structure, where
the degree of the root of the tree is same as the number of
missing digits, and level-wise it reduces to one to obtain the leaf
vertices, where each leaf at the lowest level is a valid
permutation of all the missing digits based on the clues given in
P (and the assumed valid permutation(s) in other minigrid(s) in
subsequent iterations).
As for example, the number of clues given in minigrid 3
of Figure 1(a) is 4, and the missing digits are 2, 3, 4, 6, and 7.
The proposed algorithm likes to place each of the permutations
of these missing digits in the blank locations [1, 7], [2, 7], [2, 9],
[3, 7], and [3, 8]. Here the tree structure we like to compute is
shown in Figure 10, whose root does not contain any
permutation of the missing five digits, and it is represented by
„‟. This root is having five children where the first child
leads to generate all valid permutations staring with 2, the
second child leads to generate all valid permutations staring
with 3, and so on.
Now note that none of the permutations starting with 2
is a valid permutation as column 7 of minigrid 6 contains 2 as
given clue (at location [6, 7]). So, we do not expand this vertex
(i.e., vertex with permutation „2‟) further in order to

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1146
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

compute only the set of desired valid permutations. Similarly,

we do not expand the child vertex with permutation „6‟, as
location [1, 3] contains 6 as given clue. Up to this point in time,
as either 3, or 4, or 7 could be placed at [1, 7], we expand each
of the child vertices starting with permutations 3, and 4, and 7,
as shown in Figure 10.
Similarly, we expand the tree structure inserting a new
missing number at its respective location (for a blank cell)
leading from a valid permutation (as vertex) in the previous
level of the tree. Correspondingly, we verify whether the
missing digit could be placed at the respective location for a
blank cell of the given Sudoku instance P. If the answer is „yes‟,
we further expand the vertex; otherwise, we stop expanding the
vertex in some earlier level of the tree structure prior to the last
level of leaf vertices only. As for example, the vertex with
permutation „742‟ is not expandable, because we cannot place
2 at [2, 9] as [2, 1] contains a 2 as given clue. So, this is how
either a valid permutation is generated from the root of the tree
structure reaching to a bottommost leaf vertex, or the process of
expansion is terminated in some earlier level of the tree that
must generate other than valid (unwanted) permutations at
this point in time.
Interestingly, Figure 10 shows the reality that the
number of possible permutations of five missing digits is 120,
and out of them only seven are valid for minigrid 3 of the
Sudoku instance shown in Figure 1(a). Note that the given
clues in P are nothing but constraints and we are supposed to
obey each of them. So, usually, if there are more clues, P is
more constrained and hence the number of valid permutations
is even much less, and the solution, if it exists, is unique in
most of the cases. On the contrary, if there are fewer clues in P,
more valid permutations for some minigrid of P could be
generated, computation of a solution for P might take more
time, and P may have two or more valid solutions. In any case,
if there is a unique solution of the assumed Sudoku instance (in

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1147
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

Figure 1(a)), out of these seven valid permutations only one will
finally be accepted following the subsequent steps of the
algorithm.
Now the algorithm considers one valid permutation (out
of the seven permutations) of minigrid 3 and all given clues in
P, and generates all valid permutations for minigrid 1. If at
least one valid permutation for minigrid 1 is obtained, we
proceed for generating all valid permutations for minigrid 7
obeying all given clues in P and the assumed valid
permutations of minigrids 3 and 1; otherwise, a second valid
permutation of minigrid 3 is considered, for which in a similar
way, we generate all valid permutations for minigrid 1, and so
on.
This is how the algorithm proceeds and generates all
valid permutations of a minigrid under consideration obeying
the given clues in P and a set of assumed valid permutations,
one for each of the minigrids considered earlier in succession,
up to this point in time.

Figure 10: The permutation tree for generating only valid

permutations of the missing digits in minigrid 3 of the Sudoku
instance shown in Figure 1(a).

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1148
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

Note that at the time of computing a set of valid

permutations for a minigrid, we have to consider clues and
(earlier computed) valid permutations in only four of the
remaining eight minigrids that are adjacent to the minigrid
(currently) under consideration. As for example, while
computing valid permutations for minigrid 7, we have to
consider one valid permutation of minigrid 1 and the clues
given in minigrids 1, 4, 8, and 9 only; here the assumed valid
permutation of minigrid 3 has no use while computing valid
permutations for minigrid 7. In the same way, while computing
valid permutations for minigrid 6, only we have to consider the
assumed valid permutation of minigrid 3 (up to this point in
time) and the clues given in minigrids 3, 4, 5, and 9 only; here
the assumed valid permutations of minigrids 1 and 7 have no
use while computing valid permutations for minigrid 6, and so
on.
Now we discuss about the size of the tree structure
under consideration. If p be the number of blank cells in a
minigrid and the Sudoku instance is of size nn, then the
computational time as well as the computational space
complexities of the guessed free Sudoku solver developed herein
is (p!x)n = O(pn), where x is the number of other than valid
permutations based on the clues given in the Sudoku instance
P. Our observation is that for a given Sudoku instance P, x is
very close to p!, and hence p!x is a reasonably small number
and in our case the value of n is equal to 9. Hence, the
experimentations made by this algorithm take negligible
amount of clock time, of the order of milliseconds. We conclude
that the algorithm developed in this paper guarantees a valid
solution of a given Sudoku puzzle, if one exists, and
theoretically it takes time and space exponential in size; these
has been stated in the following theorem and proved thereafter.

Theorem 2: The guessed free Sudoku solver developed

in this paper guarantees a valid solution of a Sudoku instance,

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1149
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

if one exists, and it takes time and space complexities O(pn),

where p is the number of blank cells in a minigrid of a Sudoku
instance of size nn.
Proof: The guessed free Sudoku solver developed in this paper
considers the minigrids of a given Sudoku instance P in a
particular sequence in a deterministic way. Hence, if there be a
solution S for P, then each minigrid must have its own valid
permutation for its missing digits and in combination of all of
them an overall valid solution S for P is obtained, if S is unique.
Incidentally, the guessed free Sudoku solver generates a set of
only valid permutations for a minigrid, and the required valid
permutation for the minigrid in S must be a member of this set.
If the algorithm starts from a valid permutation for a minigrid,
or considers a valid permutation for some subsequent minigrid,
which may not be a valid permutation towards computing S,
then the algorithm must reach to a point when no valid
permutation for a later minigrid will be generated, and we have
to revert back to the previous minigrid to consider its next valid
permutation.
In this process of computation, if S is unique for P, the
algorithm must consider the valid permutation of the initial
minigrid, which will be the final valid permutation for the
minigrid in S, then generation of permutations of the
subsequent minigrids and their consideration will eventually
lead to the desired solution S for P. This process might have
several backtrackings, but as the number of valid permutations
for a minigrid is negligibly less and the number of minigrids is
confined to nine only, S is guaranteed to be computed in a
reasonable amount of time, if it exists.
Now it is straightforward to prove that the algorithm
takes both computational time and space O(pn), as generation of
all valid permutations of a minigrid is the dominating
computations involved in this algorithm, where p is the number
of blank cells in a minigrid of P of size nn. Here pn is the size
of the tree structure we compute while generating all valid

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1150
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

permutations of the missing digits in a minigrid. In practice,

the size of the tree structure is significantly less than the
asymptotic upper bound mentioned herein.

4. Conclusion

In this paper we have proposed an exclusive minigrid

based Sudoku solver algorithm, which is completely guessed
free. The solver considers each of the minigrids (instead of
blank cells in isolation) of size 3×3 each that has been
developed for the first time in designing such an algorithm for a
given Sudoku puzzle of size 9×9. In our algorithm a pre-
processing is there for computing only all valid permutations
for each of the minigrids based on the clues in a given Sudoku
puzzle.
It has been observed in most practical situations that
the number of valid permutations is appreciably less than the
total number of possible permutations for each of the minigrids,
and even if there are less clues in some minigrid of an instance,
clues present in four adjacent row and column minigrids
severely help in reducing the ultimate number of valid
permutations for that minigrid too. Anyway, this approach of
minigrid-wise computation of valid permutations and checking
their compatibility among row minigrids and column minigrids
is absolutely new and done for the first time in this domain of
work.
As we consider minigrids for finding only the valid
solutions of a given Sudoku puzzle instead of considering the
individual (blank) cells, therefore, the computations involved in
the algorithm is significantly reduced. In the case of a 9×9
Sudoku puzzle, there are 81 such cells with some clues (which
is less) and the remaining blank cells (which is more), whereas
there are only nine such minigrids each of which consists of 3×3
cells. Here the observation to a Sudoku puzzle is not by
searching of missing numbers cell-by-cell (as if searching for an

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1151
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

address by moving through streets); rather, it is one step above

the ground of the puzzle by considering groups of cells or
minigrids (and searching the same from a bird‟s eye view).
The brilliancy of the algorithm developed in this paper is
that the same logic can also be straightway applied for larger
Sudoku instances such as 16×16, 25×25, or of any other
rectangular sizes (with their respective objective functions).
The level of difficulty is another important issue that
almost all the earlier Sudoku solvers consider while developing
an algorithm. There are no hard and fast rules that state the
difficulty level of a Sudoku puzzle. A sparsely filled Sudoku
puzzle may be extremely easy to solve, whereas a densely filled
Sudoku puzzle may actually be more difficult to crack. From a
programming viewpoint, we can determine the difficulty level of
a Sudoku puzzle by analyzing how much effort must be
expended to solve the puzzle, and the different levels of
difficulty are easy, medium, difficult, extremely difficult, etc.
Incidentally, the Sudoku solver developed in this paper does not
depict any level of difficulty; rather, all the Sudoku instances
are having the same level of difficulty. In some cases, more
valid permutations may be generated for some minigrid, but in
general, the number of valid permutations is much less, and the
minigrids with smaller number of valid permutations in fact
guide to compute eventually all the desired solutions for a given
Sudoku instance.

BIBLIOGRAPHY:

Chang, C.-C., P.-Y. Lin, Z. H. Wang, and M. C. Li. 2010.

“A Sudoku based Secret Image Sharing Scheme with
Reversibility.” Journal of Communications 5(1): 5-12.
Erlich, Y., K. Chang, A. Gordon, R. Ronen, O. Navon, M.
Rooks, and G. J. Hannon. 2009. “DNA Sudoku: Harnessing

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1152
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

High-throughput Sequencing for Multiplexed Specimen

Analysis.” Genome Research Journal 19(7): 1243-1253.
Hong, W., T.-S. Chen, and C.-W. Shiu. 2008.
“Steganography using Sudoku Revisited.” Proc. 2nd
International Symp. Intelligent Information Technology
Application: 935-939. Accessed June 2013, Available at
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=4739900
&url=https%3A%2F%2Ffanyv88.com%3A443%2Fhttp%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.j
sp%3Farnumber%3D4739900.
Jussien, N. 2007. A-Z of Sudoku. USA: ISTE Ltd.
Lee, W.-M. 2006. Programming Sudoku. USA: Apress.
Mantere, T. and J. Koljonen. 2007. “Solving, Rating and
Generating Sudoku Puzzles with GA.” IEEE Congress on
Evolutionary Computation: 1382-1389.
Naini, P. M., S. M. Fakhraie, and A. N. Avanaki. 2010.
“Sudoku Bit Arrangement for Combined Demosaicking and
Watermarking in Digital Camera.” Proc. 2nd International
Conf. Advances in Databases, Knowledge and Data
Applications: 41-44. Accessed June 2013, Available at
http://ieeexplore.ieee.org/xpl/articleDetails.jsp?reload=true&arn
umber=5477147&contentType=Conference+Publications.
Pacurib, J. A., G. M. M. Seno, and J. P. T. Yusiong. 2009.
“Solving Sudoku Puzzles using Improved Artificial Bee Colony
Algorithm.” In: Proc. 4th Int. Conf. Innovative Computing,
Information and Control: 885-888. Accessed July 2013,
Available at
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=5412260
&url=https%3A%2F%2Ffanyv88.com%3A443%2Fhttp%2Fieeexplore.ieee.org%2Fiel5%2F5410746
%2F5412194%2F05412260.pdf%3Farnumber%3D5412260.
Thaens, R. 2008. ”Track Maintenance through
Cooperating Agents.” Proc. 11th International Conference on
Information Fusion: 1-8. Accessed July 2013, Available at
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=4632285
&url=https%3A%2F%2Ffanyv88.com%3A443%2Fhttp%2Fieeexplore.ieee.org
%2Fxpls%2Fabs_all.jsp%3Farnumber%3D4632285.

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1153
Arnab K. Maji, Sudipta Roy, Rajat K. Pal – A Novel Algorithmic approach for
solving Sudoku puzzle in Guessed Free Manner

Wu, W.-C. and G.-R. Ren. 2009. “A New Approach to

Image Authentication using Chaotic Map and Sudoku Puzzle”,
Proc. 5th Int. Conf. Intelligent Information Hiding and
Multimedia Signal Processing: 628-631. Accessed May 2013,
Available at
http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=
5337415&contentType=Conference+Publications&queryText%3
DA+New+Approach+to+Image+Authentication+Using+Chaotic
+Map+and+Sudoku+Puzzle.
Wu, Y., J. P. Noonan, and S. Agaian. 2011. “Image
Encryption using the Rectangular Sudoku Cipher.” Proc.
International Conference on System Science and Engineering:
704-709. Accessed July 2013, Available at
http://ieeexplore.ieee.org/xpl/articleDetails.jsp?reload=true&arn
umber=5961994&contentType=Conference+Publications.
Yato, T. and T. Seta. 2003. “Complexity and
Completeness of Finding Another Solution and Its Application
to Puzzles.” IEICE Transactions on Fundamentals of
Electronics, Communications, and Computer Sciences E86-A(5):
1052–1060.

EUROPEAN ACADEMIC RESEARCH, VOL. I, ISSUE 6/ SEPTEMBER 2013

1154