Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

DEPARTMENT OF MATHEMATICS

MODELAMIENTO III
2020-I

RUBIK’S CUBE, MAGIC OR SCIENCE?


Diana Leticia Rincón Delgado
Abstrac
A formalism of all the mathematics behind the Rubik’s cube is made, characterizing by
means of theorems a part of the mathematical structure of the cube based on articles and
documents. In addition, an approach to homeomorphisms and isomorphisms is made in some
applications of cube movements, defining a group of legal movements and another of legal
movements in practice.

1 Introduction
The objective of this investigation is to know the mathematical part behind the Rubik’s cube,
from group theory, where fundamental theorems are described which allow a description of
this "game". Also, formulate homeomorphisms and isomorphisms between cube movements.

First, a study of the results or theorems about permutations is made, for example, the first
fundamental theorem of cube theory, which indicates the notation to be used to mention the
cube, such as the notation by faces and by subcubes.

Then, switches and conjugates are defined to describe what is called "The corona group"
to be able to build subgroups and combinatorics with the corners and with the edges of the
cube, and in this way to be able to define the homeomorphism between the groups with corona
products and later an isomorphism between these.

On the other hand, a study is made on a general algorithm, that is, an algorithm that only
describes mathematically, by means of theorems, the reorganization of vertices and edges.

Finally, the algorithm is described from conjugates and conjugates to obtain a better ap-
proximation of the solution algorithm.

2 Methodology
1. A theoretical structure is made, giving some definitions, and important theorems of
Group theory.

2. A formalism on Cube theory is described, with theorems and Lemmas on the cube.

3. A homomorphism and isomorphism are described from the "corona products".

4. A practice is carried out implementing the formalism in the cube.

1
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

3 Theory
3.1 Definition:[5]
Let G be a set, G is said to be a group if it is endowed by an ∗ operation and meets the
following conditions:

• a∗e=a y e?a=a

• a∗b=e y b?a=e

• a ∗ (b ∗ c) = (a ? b) ? c

• For any a, b ∈ G, a ∗ b ∈ G

3.2 Definition:
If X is a non-empty set, a permutation of X is a bijection α : X → X. The group of
permutations of X by SX will be denoted.

Notice that |Sn | = n!

3.3 Properties of permutations:[4]


• For any set A, two permutations in S(A) are called disjoint when each element of A is
a fixed point of one or the other of them.

• A cycle in S(A) is a permutation that leaves all elements fixed.

• A cycle in S(A) is a permutation that fixes all the elements except in a certain list
a1 , ..., ak , sends ai in ai+1 for 1 ≤ i < k and send ak in a1 .

• When k = 2, the cycle is called transposition[5].

3.4 Theorem:[4]
Each permutation alpha inSn is either a cycle or a product of disjoint cycles.
Proof: We will first prove the existence of decomposition and then its uniqueness. Be
 
1 2 ... n
σ=
a1 a2 . . . an

an arbitrary permutation of Sn . If σ = 1, that’s it.


Suppose σ 6= 1, and let b be a number such that σ(b) 6= b. Then σ 2 (b), σ 3 (b), ... is
constructed until b is obtained after, say, i + 1 steps. So the cycle

(b σ(b) σ 2 (b) ... σ(b))

2
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

describes part of the permutation σ. If the rest of the permutation does not contain numbers
or consists of numbers that remain invariant, the result has been demonstrated, since σ is,
then, a cycle. Otherwise, the previous procedure is repeated with that remainder and so on.
After a finite number of steps there is decomposition.
As for uniqueness, let’s note that two disjoint cycles commute. Let’s suppose

σ = σp ...σ2 .σ1 = τq ...τ2 .τ1

they were two decompositions of σ as a product of disjoint cycles. Let a1 be a number that
does not remain invariant by σ. It is evident that a1 must be in a cycle and only one of the
{σp , ..., σ2 , σ1 }, and in one and only one of the { tauq , ..., tau2 , tau1 }. By commutativity it
can be assumed that a1 is in σ1 and in τ1 . As the numbers that appear in σ1 remain invariant
for the rest of the σi cycles, the element a1 must be transformed into the same element a2 by
σ1 and by τ1 .
For the same reason, a2 must be transformed into the same element a3 using σ1 and τ1 ,
and so on. This proves that σ1 = τ1 . Repeating this reasoning, it follows that p = q and the
cycles σi are equal to τi [4]


3.5 Theorem:[4]
Every permutation α ∈ Sn is a product of transpositions.
Proof: By the previous theorem, it is enough to prove that every cycle of Sn can be
decomposed as a product of transpositions. Let σ = (a1 a2 ...am ). It’s clear that

σ = (a1 a2 ...am ) = (a1 am )...(a1 a3 )(a1 a2 )

3.6 Definition: (Semi-Direct Product)[1]


Let G1 and G2 subgroups. So A = G1 o G2 It is a semi-direct product if:

1. A = G1 G2 .

2. G1 ∩ G2 = eA Where eA is the identity element of A.

3. G1 C A

3.7 Definition: (Wreath Products)[1]


Let X be a finite set, G and H a group acting on X. Fix a labelling of X, say {x1 , x2 , ..., x4 },
with |X| = t. Let G t be the direct product of G with itself t − times. Then the wreath product
of G y H is G t o H = G t o H where H acts on G t by its action on X

3
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

3.8 Definition:[1]
The permutation group G = hF, L, U, D, R, Bi ⊂ S54 is called Grupo del cubo de Rubik

3.9 FIRST FUNDAMENTAL THEOREM OF CUBE THEORY[1]


Theorem: Let v ∈ C38 , w ∈ C212 , and s ∈ S12 . The 4 − tupla(v, r, w, s) corresponds to a
possible cube position if and only if

1. sgn(r) = sgn(s)(equal parity of permutations).

2. v1 + v2 + v3 + · · · + v8 = 0mod(3) (conservation of the total number of twists).

3. w1 + w2 + w3 + · · · + w12 = 0mod(2) (conservation of the total number of flips).

Proof:
=⇒) Let v ∈ C38 , r ∈ S8 , w ∈ C212 , s ∈ S12 and g ∈ G where g is a move that rearranges the
cube from the solve state to a state (v, r, w, s). So g can be written as g = M1 M2 ...Mn where
Mi ∈ {F, L, U, B, R, D}

1. With each move a total of four edge cubes and four corner cubes are moved; thatis, the
same number of corner cubes are moved and the same number of edge cubes are moved.
Note that each permutation is a 4−cycle, which is odd and has
n
Y
sgn = −1, so f or each g : sgn(r) = sgn(Mi ) = sgn(s)
k=1

2. Note that if Mi is U or D, then v remains unchanged, since the all corner cubes remain on
the same face. If Mi is R, L, F , or B, then two corner cubes are moved. One corner cube
is moved down off the U face and one corner cube is moved up onto the U face. So, the
components of v are either decreased by 1(mod3) or increased by 1(mod3), respectively.
But this means that for any R, L, F or B, v1 + v2 + v3 + · · · + v8 = 1(mod3) = 0(mod3).
So v1 + v2 + v3 + · · · + v8 = 0(mod3).

3. For each move g a total of four edge cubes will be reoriented. So w1 +w2 +w3 +· · ·+w12 =
4(mod2)

⇐=) Let A = (v, r, w, s) and let A satisfy conditions (1), (2), (3).

Condition (1) says that


sgn(s) = sgn(r)
So there is equal parity of permutations. Thus the permutations of the corner cubes and edge
cubes are either both even or both odd. Assume that

sgn(s) = sgn(r) = 1

4
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

that is, the permutations are even. If the permutations are odd, simply apply one of the basic
moves (B, F, L, U, R, D) and the new position will satisfy

sgn(s) = sgn(r) = 1

Now, consider the move for a corner 3−cycle. Take

M = RB −1 RF 2 R−1 BRF 2 R2

for example. The move M cycles the upper-front-left, upper-front-right, and upper-backright
corner cubes without changing the position of the other cubes. Denote the upperfront-left
cube as a1 , the upper-front-right cube as a2 , and the upper-back-right cube as a3 and denote
the remaining corner cubes as a4 , a5 , a6 , a7 , a8 . For every ai from a4 , a5 , a6 , a7 , a8 there exists
a move x from {B, F, L, U, R, D} of at most two moves (that is, two of B, F, L, U, R, D such
that ai is moved to the position of a3 without changing the position of a1 and a2 . Now, apply
the transformation xM x0 . This move creates the 3−cycle (a1 , a2 , ai ). This 3−cycle can be
obtained for any of the ai ’s; that is, the 3−cycles

(a1 , a2 , a3 ), (a1 , a2 , a4 ), (a1 , a2 , a5 ), (a1 , a2 , a6 ), (a1 , a2 , a7 ), (a1 , a2 , a8 )

can all be obtained by the appropriate move x. But, this generates all the even permutations
of the corner cubes. Therefore, there exists an appropriate move x that will return all the
corner cubes to their home positions.

Now, consider the move for an edge 3−cycle. Take the move

M ∗ = R2 U F B 0 R2 F 0 BU R2

The move M ∗ cycles the upper-front, upper-back and upperright edge cubes without chang-
ing the position of any of the other cubes. Denote the upper-front cube as b1 , the upper-
back cube as b2 and the upper-right cube as b3 and denote the remaining edge cubes by
b4 , b5 , b6 , b7 , b8 , b10 , b11 , b12 . Likewise with the corner cubes, for any bi from b4 , b5 , b6 , b7 , b8 , b10 , b11 , b12 ,
there exists a move y from {B, F, L, U, R, D} of at most 2 moves so that the edge cube bi is
moved to the position of b3 without changing the position of b1 and b2 . Now, apply the trans-
formation yM ∗ y 0 .

This transformation creates the 3−cycle (b1 , b2 , bi ). But using an appropriate choice for y,
the 3−cycles
(b1 , b2 , b3 ), (b1 , b2 , b4 ), (b1 , b2 , b5 ), (b1 , b2 , b6 ), (b1 , b2 , b7 ),
(b1 , b2 , b8 ), (b1 , b2 , b9 ), (b1 , b2 , b10 ), (b1 , b2 , b11 ), (b1 , b2 , b12
These generate all the even permutations of the edge cubes. Thus, there exists an appropriate
move y that will return all of the edge cubes into their home positions.

5
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

All that is left to do is to reorient the cubes so that the facets are colour matched. Con-
dition (2) says that there is a conservation of total twists; that is, the number of clockwise
twists is equal to the number of counterclockwise twists. This means there exists a move which
twists exactly 2 corner cubes and preserves the orientation and position of all the other cubes,
namely the move
M1 = (R0 D2 RB 0 U 2 B 2 )
which twists the upper-front-right corner cube by 120 degrees and twists the bottom-down-left
cube by −120 degrees. Note that the move M1 can be modified to obtain a similar result for any
2 corner cubes. To begin to match the facets of the corner cubes, first twist any clockwise and
counterclockwise pairs into their solved orientations. The remaining corner cube orientations
will occur in triples since the corner cubes obey
8
X
vi = 0(mod3)
i=1

So they will occur in either 3 clockwise twists or 3 counterclockwise twists. Call these 3 cubes
c1 , c2 , c3 .The remaining corner cubes can be solved by a sequence of corner twisting moves,
say
M1∗ = L0 D2 LBD2 B 0 U BD2 B 0 L0 D2 LU 0
or a similar move for two of the remaining corner cubes that need reorienting. Now, M1∗ will
solve one of the remaining corner cubes, say c1 , and reorient the other corner cube, say c2 , into
the opposite position to the untouched corner cube, c3 . That is, if c3 needs to be solved by a
clockwise twist then M1∗ will reorient c2 to a position that needs a counterclockwise twist to
be solved and vice versa. The remaining two cubes can be solved with the appropriate move
M1 . Thus all the corner cubes are in their solved states.

Condition (3) says that there is a conservation of total flips. Since


12
X
wi = 0(mod2)
i=1

, there is an even number of edge cubes that need to be flipped. But there exists a move that
flips exactly 2 edge cubes and preserves the orientation and position of the remaining cubes.
Take the move
M2 = LF R0 F 0 L0 U 2 RU RU 0 R2 U 2 R
The move M2 flips the upper-front edge cube and the upper-right edges cube, while leaving
the position and orientation off all the other cubes untouched. The move M2 can be modified
appropriately so that any 2 edges cubes can be flipped and the position and orientation of all
the other cubes will be preserved. Since there are an even number of edge cubes, all the edge
cubes are able to return to their solved orientations.

6
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

Thus A is a solvable position on the Rubik’s Cube. So (v, r, w, s) is a possible arrangement of


the Rubik’s Cube[1].


3.10 SECOND FUNDAMENTAL THEOREM OF CUBE THEORY[1]


Theorem: An operation of the cube is possible if and only if the following are satisfied:
1. The total number of edge and corner cycles of even length is even

2. The number of corner cycles twisted right is equal to the number of corner cycles twisted
left (up to modulo 3).

3. There is an even number of reorienting edge cycles.


Proof =⇒) Let M be an operation on the cube that takes the cube from the solved state
to position g = (v, s, w, r), where v ∈ C38 , r ∈ S8 , w ∈ C212 , s ∈ S12
1. By (1) of (F irstT heorem), sgn(r) = sgn(s). But this means that the permutation is
even. So the length of the edge and corner cycles is even.

2. For any move M , the corner cubes are moved either right, left or not at all. So the cycle
changes the sum of vi by 2, 1 or 0(mod3) respectively. By (F irstT heorem)
8
X
vi = 0
i=1

the number of right twists is equal to the number of left twists.

3. Note that an edge cycle only reorients if it is changed by an odd number; that is wj = 1
for some j = {1, 2, 3, ..., 12}. By (F irstT heorem)
12
X
wi = 0
i=1

But this means that if one edge cycle is reorienting, then another edge cycle must be
reorienting since the sum is zero. Thus, there must be an even number of reorienting
edge cycles.
⇐=) Suppose that (1), (2), and (3) hold. By (F irstT heorem), there exists move M that takes
the cube from the the solved state to the state g. There also exists move M 0 that takes the
cube from the state g to the solved state. Now, by assumption M and M 0 satisfy (1), (2), and
(3). But M and M 0 are both valid operations on the Rubik’s Cube. Thus, if (1), (2), and (3)
hold, then the operation is valid[1].


7
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

3.11 LEGAL AND ILLEGAL GROUP[1]


Using the first and second fundamental theorem of cube theory, the group of illegal Rubik’s
cubes can be reduced to the group

G0 = {(v, r, w, s) : v ∈ C812 , r ∈ S8 , w ∈ C212 , s ∈ S12 }

Notice that G0 satisfies the first and second fundamental theorems.

3.12 Lemma1:[1]
Let S1 , S2 , ..., Sn finite sets. So

|S1 × S2 × · · · × Sn | = |S1 | · |S2 | · · · · · |Sn |

Proof: Let S1 × · · · × Sk = {(s1 , ..., sn ) : si ∈ Si }. Now, there are |S1 | choices for s1 , |S2 |
choices for s2 , |S3 | choices for s3 , and so on. By the multiplication principle[1]:

|S1 × S2 × · · · × Sn | = |S1 | · |S2 | · · · · · |Sn |

3.13 Theorem:[1]
There is an isomorphism:
G0 ∼
= (C37 o S8 ) × (C211 o S12 )
and
|G0 | = |S8 ||S12 ||C211 ||C37 | = 8! · 12! · 211 · 37
Proof: By (F irstT heorem), the First Isomorphism Theorem of Groups and the definition
of semi-direct product
G0 ∼
= (C37 o S8 ) × (C211 o S12 )
By Lemma1[1]
|G0 | = |S8 ||S12 ||C211 ||C37 | = 8! · 12! · 211 · 37


3.14 Lemma2:[1]
The Rubik’s Cube Group, G, can be expressed as

G = (C37 o S8 ) × (C210 o S12 )

Proof: This follows from Lemma1 and F irstT heorem[1]




8
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

3.15 Corollary:[1]
The Rubik’s Cube Group G is the kernel of the homomorphism
φ : G0 → {1, −1}
(v, r, w, s) → sgn(r)sgn(s)
In particular G ⊂ G0 is normal of index 2 and
|G| = 8! · 12! · 210 · 37
Proof: Let G0 = (C37 o S8 ) × (C211 o S12 ), H = {−1, 1} y φ : G0 → H, so
ker(φ) = {(v, r, w, s) : φ(v, r, w, s) = eH }, eH = 1
By first fundamental theorem of the first group isomorphism theorem we have to G0 /ker(φ) ∼ =
G, with G = (C37 o S8 ) × (C210 o S12 ), then for the previous motto |G| = 8! · 12! · 210 · 37 and [1]
(8! · 12! · 211 · 37 )
[G0 : G] = =2
(8! · 12! · 210 · 37 )


3.16 Switches[2]
Let X and Y be two moves in the Rubik’s cube, then it is possible that:
XY 6= Y X
In this case we say that X and Y do not commute.

For example: F U 6= U F .

But some pair is of movements if they commute, for example: U D = DU


• If X and Y are two movements in the Rubik’s cube, their switch is the movement
XY X 0 Y 0
• XY X 0 Y 0 = 1, so X and Y are two movements that commute, since XY X 0 Y 0 =
(XY )(Y X)0 , is equal to the identity when (Y X)0 is the inverse of XY , that is when
XY = Y X

3.17 CONJUGATION[2]
Conjugated movements do the same kind of thing, but in different places.

For example:
• If X flips two edges, then ZXZ 0 flips two edges (generally different)
• If X switches three edges, then ZXZ 0 switches three (usually different) edges.

9
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

4 Practice
4.1 NOTATION:
The notation used to represent the cube is the notation for faces, that is: Front (F ), Up (U ),
Right (R), Left (L), Down (D), Back (B), as you can see in figure 1.

Figure 1:

4.2 Cube Division


Particularly, the cube is divided into corners and edges, which will be the object of the study,
because the center is invariant and always remains fixed, in addition to being a reference point.

4.2.1 Cube corners[1]


Since a corner of the cube has 3 orientations, and since there are eight corner cubes, the set
can describe the orientation of any facet of a corner of the cube.

C3 × C3 × C3 × C3 × C3 × C3 × C3 × C3 = C38

Therefore, possible cube corner arrangements can be described by the eight-element permuta-
tion group. S8

EXAMPLE:

Consider the top edge of the cube on the front face of the Rubik’s Cube. It starts with a
number of 1. Now, when you apply the F R move, the facet moves to the top face on the right
side. This edge cube position is assigned the number 0.

With each turn, the orientation number of the edge cube is changed to 0mod2 or 1mod2.

10
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

Figure 2:

4.2.2 Cube edges[1]


Since a cube edge has 2 orientations, and since there are 12 cube edges, the set can describe
the orientation of any facet of an edge.

C2 × C2 × C2 × C2 × C2 × C2 × C2 × C2 × C2 × C2 × C2 × C2 = C212

Therefore, the possible arrangements of the edges of the cube can be described by means of
the permutation group of twelve elements. S12

EXAMPLE:

Consider the cube and the top, front and left corner Cube.

With each turn of the face R, L, F or B, the orientation number of the corner facet is
changed to 1mod3 or 2mod3. With each turn of the face U or D, the numbering remains
unchanged (0mod3).

Figure 3:

4.3 Switches
4.3.1 switch two edges
1. X = LU 0 L2 U 2 L

2. Y = U

11
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

3. X 0 = L0 U 2 L2 U L0

4. Y 0 = U 0
So the switch,
XY X 0 Y 0 = (LU 0 L2 U 2 L)U (L0 U 2 L2 U L0 )U 0

This switch flips the top left and top front edges without disturbing the rest of the cube.

Figure 4: Flip two edges

4.3.2 switch three edges


• X=R

• Y = D2

• X 0 = R0

• Y 0 = D2
XY X 0 Y 0 = RD2 R0 D2

Figure 5: Toggle Edges

12
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

4.3.3 Rotate two corners


1. X = F 0 DF LDL0

2. Y = U

3. X 0 = LD0 L0 F 0 D0 F

4. Y 0 = U 0 So the switch,

XY X 0 Y 0 = (F 0 DF LDL0 )U (LD0 L0 F 0 D0 F )U 0

Rotate the top left front corner clockwise and the top right front corner counterclockwise
without disturbing the rest of the cube.

Figure 6: Corners Rotation

4.3.4 Rotate three corners


• X = LDL0

• Y =U

• X 0 = LD0 L0

• Y 0 = U0
XY X 0 Y 0 = (LDL0 )U (LD0 L0 )U 0

Figure 7: Switching Corners

13
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

4.4 Conjugation
4.4.1 conjugation three edges
Let

X =RU 2 R0 U 2
Z =F 2 U
Z 0 =U 0 F 2

So,

ZXZ 0 =(F 2 U )RU 2 R0 U 2 (U 0 F 2 )


=F 2 U RU 2 R0 U F 2

Figure 8: Toggle Edges

4.4.2 Conjugation top three corners


Let,

X =F LF 0 ; X 0 = F L0 F 0
Y =R2
Z =F 2

So, the corner switching is given by Z(XY X 0 Y 0 )Z 0

Z(XY X 0 Y 0 )Z 0 = F 0 LF 0 R2 F L0 F 0 R2 F 2

14
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

Figure 9: Swithches Corners

5 SOLUTION BY SWITCHES
Let the conjugation X 0 = Y 0 XY and the switch [X, Y ] = XY X 0 Y 0 , for X, Y cube movements.
Let MR denote clockwise (with respect to the right side).The switch method solution strategy
consists of 3 stages:

5.1 Stage 1:
Solve the top face and top edges.

In this case, since the entire cube is disordered, there is no general step in it, since what
is done is to assemble the first face regardless of the others.

15
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

5.2 Stage 2:
Work out the middle edges (and the bottom edges as best as possible).

Figure 10: L : U 0 L0 U LU F U 0 F 0 , R : U RU 0 R0 U 0 F 0 U F

5.3 Stage 3:
Work out the bottom corners (and bottom edges if necessary).

Figure 11: C : F RU R0 U 0 F 0 ; OE : U RU 0 L0 U R0 U 0 L; F : (R0 D0 RD)n

16
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

6 Conclusions
1. The Rubik’s cube was described through cube theory using the fundamental theorems of
that theory. In addition, permutations were defined from the group theory to describe
the movements of the cube.

2. Switches and conjugates were defined, describing a very simple and simplified solution.

3. The construction of the crown groups was made from the combinations of both edges and
corners defining isomorphisms between the movements, this part was important since it
allows eliminating the movements that are not possible in the cube, that is, it takes the
illegal group as a group outside the cube.

17
DEPARTMENT OF MATHEMATICS
MODELAMIENTO III
2020-I

References
[1] Daniels-Project/Rubik’s cube.pdf

[2] Mathematics Rubik’s cube.pdf

[3] Rubik’s cube.pdf

[4] Chen, Janet. ”Group Theory and the Rubik’s Cube.”

[5] The Mathematics of the Rubik’s Cube,Introduction to Group Theory and Permutation
Puzzles

[6] Alex Wein, Mathematics of the Rubik’s Cube

[7] Ramón E. Romero, The mathematics of Rubik’s cube

[8] R. Michael Porter, El álgebra del cubo de Rubik

[9] rubiks-cube-solver.com

18

You might also like