Some algebraic approaches to graceful labellings
Transcript of Some algebraic approaches to graceful labellings
Some algebraic approaches to graceful labellings
Andrea Vietri1
1Dipartimento di Scienze di Base e Applicate per l’IngegneriaSapienza Università di Roma
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 1 / 50
Outline
1 Graceful labelings
2 Graceful polynomials
3 Vanishing polynomials of small degree
4 Final remarks on graceful polynomials
5 Real-graceful labellings
6 Graceful hypersurfaces
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 2 / 50
Graceful labelings
Graceful labellings
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 3 / 50
Graceful labelings
A graph with e edges has a
graceful labelling
if we can assign positive integers in [0,e] to its vertices, so that thedifferences on the edges be 1,2, ...,e.
r rrr06 4
1@@��r r r r r4 0 3 1 2
r r r r r r r r r r r rr r r r r r r r r r r r17616715814913101211
519420321222123024
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 4 / 50
Graceful labelings
Graph decomposition: K13 can be decomposed into 14 copies of K4starting with a G.L. of K4 and replicating it:
r rrr06 4
1@@��
r rrr17 5
2@@�� . . .
r rrr11
4
2
12@@��
r rrr12
5
3
0@@��
counting (mod 13).
Every edge {a,b} of K13 is covered with norepetition.e.g. if b− a = ±4 (as for {1,5} or {2,11}), we choose one of the upperedges.Similarly with the differences ±1, ±2, ±3, ±5, ±6 .
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 5 / 50
Graceful labelings
Graph decomposition: K13 can be decomposed into 14 copies of K4starting with a G.L. of K4 and replicating it:
r rrr06 4
1@@��
r rrr17 5
2@@�� . . .
r rrr11
4
2
12@@��
r rrr12
5
3
0@@��
counting (mod 13). Every edge {a,b} of K13 is covered with norepetition.e.g. if b− a = ±4 (as for {1,5} or {2,11}), we choose one of the upperedges.Similarly with the differences ±1, ±2, ±3, ±5, ±6 .
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 5 / 50
Graceful labelings
A G.L. can be also regarded as a well done optimisation.
For, the best we can do in some cases is...
s ssss
0
10
8
1
4
@@@ ���
[0,10] \ {5} t {4}
oppures ssss
0
11
9
1
4
@@@ ���
[0,11] \ {6}
(cfr. F. Van Bussel, Relaxed graceful labellings of trees, Electr. J.Comb. (2002); see also the problem of Golomb rulers.)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 6 / 50
Graceful labelings
A G.L. can be also regarded as a well done optimisation.For, the best we can do in some cases is...
s ssss
0
10
8
1
4
@@@ ���
[0,10] \ {5} t {4}
oppures ssss
0
11
9
1
4
@@@ ���
[0,11] \ {6}
(cfr. F. Van Bussel, Relaxed graceful labellings of trees, Electr. J.Comb. (2002); see also the problem of Golomb rulers.)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 6 / 50
Graceful labelings
A G.L. can be also regarded as a well done optimisation.For, the best we can do in some cases is...
s ssss
0
10
8
1
4
@@@ ���
[0,10] \ {5} t {4}
oppures ssss
0
11
9
1
4
@@@ ���
[0,11] \ {6}
(cfr. F. Van Bussel, Relaxed graceful labellings of trees, Electr. J.Comb. (2002); see also the problem of Golomb rulers.)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 6 / 50
Graceful labelings
As to K5, K6, ... gracefulness is lost (only relaxations hold).
Similarly for cycles of length 1 o 2 (mod 4) (Rosa’s Theorem: anEulerian graph having 1 or 2 edges (mod 4) is not graceful).
Rosa’s Theorem is an isolated example of happy marriage betweenalgebra and combinatorics in this area.
In order to celebrate the Golden Wedding of that marriage (2017), witha little delay we propose a generalisation of such a basic example!
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 7 / 50
Graceful labelings
As to K5, K6, ... gracefulness is lost (only relaxations hold).
Similarly for cycles of length 1 o 2 (mod 4) (Rosa’s Theorem: anEulerian graph having 1 or 2 edges (mod 4) is not graceful).
Rosa’s Theorem is an isolated example of happy marriage betweenalgebra and combinatorics in this area.
In order to celebrate the Golden Wedding of that marriage (2017), witha little delay we propose a generalisation of such a basic example!
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 7 / 50
Graceful labelings
As to K5, K6, ... gracefulness is lost (only relaxations hold).
Similarly for cycles of length 1 o 2 (mod 4) (Rosa’s Theorem: anEulerian graph having 1 or 2 edges (mod 4) is not graceful).
Rosa’s Theorem is an isolated example of happy marriage betweenalgebra and combinatorics in this area.
In order to celebrate the Golden Wedding of that marriage (2017), witha little delay we propose a generalisation of such a basic example!
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 7 / 50
Graceful labelings
As to K5, K6, ... gracefulness is lost (only relaxations hold).
Similarly for cycles of length 1 o 2 (mod 4) (Rosa’s Theorem: anEulerian graph having 1 or 2 edges (mod 4) is not graceful).
Rosa’s Theorem is an isolated example of happy marriage betweenalgebra and combinatorics in this area.
In order to celebrate the Golden Wedding of that marriage (2017), witha little delay we propose a generalisation of such a basic example!
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 7 / 50
Graceful polynomials
Graceful polynomials
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 8 / 50
Graceful polynomials
Graceful polynomials
For any graph we define a family {Sn}n∈N∗ of graceful polynomials; Snhas degree n.
There is a deep connection between these polynomials and gracefullabellings.
As mentioned, the prototype S1 was employed by A. Rosa (1967) toobtain the most important class of non-graceful graphs.
A.V. (2012, 2016): generalisation to every degree (analysis of gracefultrees; construction of non-graceful graphs).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 9 / 50
Graceful polynomials
Graceful polynomials
For any graph we define a family {Sn}n∈N∗ of graceful polynomials; Snhas degree n.
There is a deep connection between these polynomials and gracefullabellings.
As mentioned, the prototype S1 was employed by A. Rosa (1967) toobtain the most important class of non-graceful graphs.
A.V. (2012, 2016): generalisation to every degree (analysis of gracefultrees; construction of non-graceful graphs).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 9 / 50
Graceful polynomials
Graceful polynomials
For any graph we define a family {Sn}n∈N∗ of graceful polynomials; Snhas degree n.
There is a deep connection between these polynomials and gracefullabellings.
As mentioned, the prototype S1 was employed by A. Rosa (1967) toobtain the most important class of non-graceful graphs.
A.V. (2012, 2016): generalisation to every degree (analysis of gracefultrees; construction of non-graceful graphs).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 9 / 50
Graceful polynomials
Graceful polynomials
For any graph we define a family {Sn}n∈N∗ of graceful polynomials; Snhas degree n.
There is a deep connection between these polynomials and gracefullabellings.
As mentioned, the prototype S1 was employed by A. Rosa (1967) toobtain the most important class of non-graceful graphs.
A.V. (2012, 2016): generalisation to every degree (analysis of gracefultrees; construction of non-graceful graphs).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 9 / 50
Graceful polynomials
Graceful labellings
A graph G = (V ,E) is graceful if
∃ f : V → {0,1,2, ..., |E |} injective and such that
{|f (u)− f (v)| : uv ∈ E} = {1,2, ..., |E |} .
If a graph admits no such labelling, it is non-graceful.
tttt tt t
2
8
7
0
1
4
35
6 12
3
4
7
8@@@ t t
t t���@@@
4
6
1
0
ttt tt
Non-graceful
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 10 / 50
Graceful polynomials
Example of graceful polynomial
How to obtain a graceful polynomial from G = rr rr rr r@@ :
Assign a variable to each vertex: rr rr rr rx4
x1
x5
x2
x6
x3
x7
@@
Fix a positive integer n, say n = 3 ...
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 11 / 50
Graceful polynomials
Example of graceful polynomial
How to obtain a graceful polynomial from G = rr rr rr r@@ :
Assign a variable to each vertex: rr rr rr rx4
x1
x5
x2
x6
x3
x7
@@
Fix a positive integer n, say n = 3 ...
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 11 / 50
Graceful polynomials
Example of graceful polynomial
How to obtain a graceful polynomial from G = rr rr rr r@@ :
Assign a variable to each vertex: rr rr rr rx4
x1
x5
x2
x6
x3
x7
@@
Fix a positive integer n, say n = 3 ...
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 11 / 50
Graceful polynomials
Take any set of 3 edges
u uu u ur rx1
x5
x2
x6 x7
@@ and store :
(x1 − x2)(x2 − x5)(x6 − x7) ≡ x1x2x6 + x1x2x7 + ... (mod 2)
SUM UP these polynomials over all the(8
3
)sets of 3 edges:
this is S3(G), the 3-rd graceful polynomial of G.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 12 / 50
Graceful polynomials
Take any set of 3 edges
u uu u ur rx1
x5
x2
x6 x7
@@ and store :
(x1 − x2)(x2 − x5)(x6 − x7) ≡ x1x2x6 + x1x2x7 + ... (mod 2)
SUM UP these polynomials over all the(8
3
)sets of 3 edges:
this is S3(G), the 3-rd graceful polynomial of G.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 12 / 50
Graceful polynomials
Formally... given a graph G...
Assign variable xi to vertex vi .
Associate any given edge ej = vpvq to the polynomial Pj = xp + xq.
The n-th graceful polynomial of G is
SnG(x1, x2, ..., x|V |) ≡
∑1≤j1<j2<...<jn≤|E |
Pj1Pj2 · · · Pjn (mod 2).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 13 / 50
Graceful polynomials
Formally... given a graph G...
Assign variable xi to vertex vi .
Associate any given edge ej = vpvq to the polynomial Pj = xp + xq.
The n-th graceful polynomial of G is
SnG(x1, x2, ..., x|V |) ≡
∑1≤j1<j2<...<jn≤|E |
Pj1Pj2 · · · Pjn (mod 2).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 13 / 50
Graceful polynomials
Formally... given a graph G...
Assign variable xi to vertex vi .
Associate any given edge ej = vpvq to the polynomial Pj = xp + xq.
The n-th graceful polynomial of G is
SnG(x1, x2, ..., x|V |) ≡
∑1≤j1<j2<...<jn≤|E |
Pj1Pj2 · · · Pjn (mod 2).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 13 / 50
Graceful polynomials
Graceful polynomials and non-graceful graphs
The following Lemma provides the connection:
Lemma (A.V. 2012)Let G be a graceful graph and let fi be the label of vertex vi . Then, forevery positive integer n ≤ |E(G)|,
SnG(f1, ..., f|V |) ≡
([ |E |+1
2 ]
n
)(mod 2).
Therefore, if some graceful polynomial VANISHES (mod 2)...... while |E | makes the binomial coefficient ODD...
... then we have a contradiction,and the graph is necessarily NON-GRACEFUL.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 14 / 50
Graceful polynomials
Graceful polynomials and non-graceful graphs
The following Lemma provides the connection:
Lemma (A.V. 2012)Let G be a graceful graph and let fi be the label of vertex vi . Then, forevery positive integer n ≤ |E(G)|,
SnG(f1, ..., f|V |) ≡
([ |E |+1
2 ]
n
)(mod 2).
Therefore, if some graceful polynomial VANISHES (mod 2)...
... while |E | makes the binomial coefficient ODD...... then we have a contradiction,and the graph is necessarily NON-GRACEFUL.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 14 / 50
Graceful polynomials
Graceful polynomials and non-graceful graphs
The following Lemma provides the connection:
Lemma (A.V. 2012)Let G be a graceful graph and let fi be the label of vertex vi . Then, forevery positive integer n ≤ |E(G)|,
SnG(f1, ..., f|V |) ≡
([ |E |+1
2 ]
n
)(mod 2).
Therefore, if some graceful polynomial VANISHES (mod 2)...... while |E | makes the binomial coefficient ODD...
... then we have a contradiction,and the graph is necessarily NON-GRACEFUL.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 14 / 50
Graceful polynomials
Graceful polynomials and non-graceful graphs
The following Lemma provides the connection:
Lemma (A.V. 2012)Let G be a graceful graph and let fi be the label of vertex vi . Then, forevery positive integer n ≤ |E(G)|,
SnG(f1, ..., f|V |) ≡
([ |E |+1
2 ]
n
)(mod 2).
Therefore, if some graceful polynomial VANISHES (mod 2)...... while |E | makes the binomial coefficient ODD...
... then we have a contradiction,and the graph is necessarily NON-GRACEFUL.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 14 / 50
Vanishing polynomials of small degree
Vanishing polynomialsof small degree
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 15 / 50
Vanishing polynomials of small degree
Main problem:
Choose n, and characterise all graphs for which SnG vanishes (mod 2).
The problem seems interesting on its own right!
However, once we have these graphs, we can also find
non-graceful graphs
(choosing a suitable number of edges, if we can...).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 16 / 50
Vanishing polynomials of small degree
Main problem:
Choose n, and characterise all graphs for which SnG vanishes (mod 2).
The problem seems interesting on its own right!
However, once we have these graphs, we can also find
non-graceful graphs
(choosing a suitable number of edges, if we can...).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 16 / 50
Vanishing polynomials of small degree
Main problem:
Choose n, and characterise all graphs for which SnG vanishes (mod 2).
The problem seems interesting on its own right!
However, once we have these graphs, we can also find
non-graceful graphs
(choosing a suitable number of edges, if we can...).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 16 / 50
Vanishing polynomials of small degree
Main problem, degree 1
This is Rosa’s polynomial.
(A. Rosa, On certain valuations of the vertices of a graph, Theory ofGraphs, Internat. Sympos. Rome, 1966, pp. 349-355)
Let δp denote the degree of vertex vp.
S1G(x1, x2, ..., x|V |) ≡
∑1≤j≤|E |
Pj ≡∑
vpvq edge
(xp + xq) ≡∑
vp vertex
δpxp
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 17 / 50
Vanishing polynomials of small degree
Main problem, degree 1
This is Rosa’s polynomial.
(A. Rosa, On certain valuations of the vertices of a graph, Theory ofGraphs, Internat. Sympos. Rome, 1966, pp. 349-355)
Let δp denote the degree of vertex vp.
S1G(x1, x2, ..., x|V |) ≡
∑1≤j≤|E |
Pj ≡∑
vpvq edge
(xp + xq) ≡∑
vp vertex
δpxp
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 17 / 50
Vanishing polynomials of small degree
∑δpxp vanishes if all degrees are even (Eulerian graph).
On the other hand,([ |E|+1
2 ]
1
)is odd if |E | ≡ 1,2 (mod 4). Therefore:
Theorem (Rosa)
Eulerian graphs with |E | ≡ 1,2 (mod 4) are non-graceful.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 18 / 50
Vanishing polynomials of small degree
∑δpxp vanishes if all degrees are even (Eulerian graph).
On the other hand,([ |E|+1
2 ]
1
)is odd if |E | ≡ 1,2 (mod 4). Therefore:
Theorem (Rosa)
Eulerian graphs with |E | ≡ 1,2 (mod 4) are non-graceful.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 18 / 50
Vanishing polynomials of small degree
∑δpxp vanishes if all degrees are even (Eulerian graph).
On the other hand,([ |E|+1
2 ]
1
)is odd if |E | ≡ 1,2 (mod 4). Therefore:
Theorem (Rosa)
Eulerian graphs with |E | ≡ 1,2 (mod 4) are non-graceful.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 18 / 50
Vanishing polynomials of small degree
Degree 2
Denote coefficients of graceful polynomials byH······ .
S2G ≡
∑H2
px2p +
∑H11
pqxpxq .
H2p counts the PAIRS of edges containing vp: s��AA
XXAA@@
vavbr r
vp(xp + xa)(xp + xb)
= x2p + ...
Therefore, H2p ≡
(δp2
)(mod 2)
⇒ δp ≡ 0,1 (mod 4) ∀vp if we ask for vanishing.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 19 / 50
Vanishing polynomials of small degree
Degree 2
Denote coefficients of graceful polynomials byH······ .
S2G ≡
∑H2
px2p +
∑H11
pqxpxq .
H2p counts the PAIRS of edges containing vp: s��AA
XXAA@@
vavbr r
vp(xp + xa)(xp + xb)
= x2p + ...
Therefore, H2p ≡
(δp2
)(mod 2)
⇒ δp ≡ 0,1 (mod 4) ∀vp if we ask for vanishing.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 19 / 50
Vanishing polynomials of small degree
Degree 2
Denote coefficients of graceful polynomials byH······ .
S2G ≡
∑H2
px2p +
∑H11
pqxpxq .
H2p counts the PAIRS of edges containing vp: s��AA
XXAA@@
vavbr r
vp(xp + xa)(xp + xb)
= x2p + ...
Therefore, H2p ≡
(δp2
)(mod 2)
⇒ δp ≡ 0,1 (mod 4) ∀vp if we ask for vanishing.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 19 / 50
Vanishing polynomials of small degree
As to H11pq , there are two cases.
r rG21
vp vq@@�����
���
PP@@
δpδq choices
AAA
��
r rG22
vp vq@@�����
���
PP@@
δpδq − 1 choices:
(xp + xq)(xp + xq) not allowed
AAA
��
So we have (if we ask for vanishing):
δp ≡ 0,1 ∀vp , (δp, δq) 6≡ (1,1)∀G21 , (δp, δq) ≡ (1,1)∀G2
2
all (mod 4)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 20 / 50
Vanishing polynomials of small degree
As to H11pq , there are two cases.
r rG21
vp vq@@�����
���
PP@@
δpδq choices
AAA
�� r rG22
vp vq@@�����
���
PP@@
δpδq − 1 choices:
(xp + xq)(xp + xq) not allowed
AAA
��
So we have (if we ask for vanishing):
δp ≡ 0,1 ∀vp , (δp, δq) 6≡ (1,1)∀G21 , (δp, δq) ≡ (1,1)∀G2
2
all (mod 4)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 20 / 50
Vanishing polynomials of small degree
As to H11pq , there are two cases.
r rG21
vp vq@@�����
���
PP@@
δpδq choices
AAA
�� r rG22
vp vq@@�����
���
PP@@
δpδq − 1 choices:
(xp + xq)(xp + xq) not allowed
AAA
��
So we have (if we ask for vanishing):
δp ≡ 0,1 ∀vp , (δp, δq) 6≡ (1,1)∀G21 , (δp, δq) ≡ (1,1)∀G2
2
all (mod 4)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 20 / 50
Vanishing polynomials of small degree
From the above conditions we obtain all possible graphs:
Theorem (A.V. 2012)
S2G vanishes for complete graphs on 4d + 2 vertices, for any integer d.
No other graph satisfies the requirement.
By extending Rosa’s counting technique to this level we obtain:
Theorem (A.V. 2012, new proof of an old result ! )Complete graphs on either 16u + 10 or 16u + 14 vertices, with anyinteger u, are non-graceful.
(count the edges and reach a contradiction through the above Lemma)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 21 / 50
Vanishing polynomials of small degree
From the above conditions we obtain all possible graphs:
Theorem (A.V. 2012)
S2G vanishes for complete graphs on 4d + 2 vertices, for any integer d.
No other graph satisfies the requirement.
By extending Rosa’s counting technique to this level we obtain:
Theorem (A.V. 2012, new proof of an old result ! )Complete graphs on either 16u + 10 or 16u + 14 vertices, with anyinteger u, are non-graceful.
(count the edges and reach a contradiction through the above Lemma)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 21 / 50
Vanishing polynomials of small degree
Degree 3
S3G ≡
∑H3
px3p +
∑H21
p,qx2p xq +
∑H111
pqr xpxqxr
(commas separate SETS of indices)
H3p ≡
(δp3
)H21
p,q(G21) ≡
(δp2
)δq , H21
p,q(G22) ≡
(δp2
)δq − (δp − 1)
r rG21
vp vq@@
���
���
PP@@
AAA
���� r rG2
2
vp vq@@�����
���
PP@@
��
rva
(xp + xq)2(xp + xa) not allowed
AAA
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 22 / 50
Vanishing polynomials of small degree
Degree 3
S3G ≡
∑H3
px3p +
∑H21
p,qx2p xq +
∑H111
pqr xpxqxr
(commas separate SETS of indices)
H3p ≡
(δp3
)
H21p,q(G2
1) ≡(δp
2
)δq , H21
p,q(G22) ≡
(δp2
)δq − (δp − 1)
r rG21
vp vq@@
���
���
PP@@
AAA
���� r rG2
2
vp vq@@�����
���
PP@@
��
rva
(xp + xq)2(xp + xa) not allowed
AAA
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 22 / 50
Vanishing polynomials of small degree
Degree 3
S3G ≡
∑H3
px3p +
∑H21
p,qx2p xq +
∑H111
pqr xpxqxr
(commas separate SETS of indices)
H3p ≡
(δp3
)H21
p,q(G21) ≡
(δp2
)δq , H21
p,q(G22) ≡
(δp2
)δq − (δp − 1)
r rG21
vp vq@@
���
���
PP@@
AAA
���� r rG2
2
vp vq@@�����
���
PP@@
��
rva
(xp + xq)2(xp + xa) not allowed
AAA
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 22 / 50
Vanishing polynomials of small degree
As to H111pqr , there are 4 subgraphs on 3 vertices:
r rr
G31
vq vr
vp r rr
vq vr
vp
G32
r rr
G33
vq vr
vp r rr
G34
vq vr
vp
For example, H111pqr (G3
3) ≡ δpδqδr − δr − δq .
(forbidden repetitions of either vpvq or vpvr ) r rr
vq vr
vp
HHH
���
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 23 / 50
Vanishing polynomials of small degree
As to H111pqr , there are 4 subgraphs on 3 vertices:
r rr
G31
vq vr
vp r rr
vq vr
vp
G32
r rr
G33
vq vr
vp r rr
G34
vq vr
vp
For example, H111pqr (G3
3) ≡ δpδqδr − δr − δq .
(forbidden repetitions of either vpvq or vpvr ) r rr
vq vr
vp
HHH
���
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 23 / 50
Vanishing polynomials of small degree
System of constraints:
∀vp δp 6≡ 3 (mod 4)
∀G21
(δp2
)δq +
(δq2
)δp ≡ 0 (mod 2) (H21
p,q +H21q,p ...)
∀G22
(δp2
)δq +
(δq2
)δp + δp + δq ≡ 0 (mod 2)
∀G31 δpδqδr ≡ 0 ...
∀G32 δp(δqδr + 1) ≡ 0
∀G33 δpδqδr + δq + δr ≡ 0
∀G34 δpδqδr + δp + δq + δr ≡ 0
Eulerian graphs clearly satisfy this system.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 24 / 50
Vanishing polynomials of small degree
System of constraints:
∀vp δp 6≡ 3 (mod 4)
∀G21
(δp2
)δq +
(δq2
)δp ≡ 0 (mod 2) (H21
p,q +H21q,p ...)
∀G22
(δp2
)δq +
(δq2
)δp + δp + δq ≡ 0 (mod 2)
∀G31 δpδqδr ≡ 0 ...
∀G32 δp(δqδr + 1) ≡ 0
∀G33 δpδqδr + δq + δr ≡ 0
∀G34 δpδqδr + δp + δq + δr ≡ 0
Eulerian graphs clearly satisfy this system.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 24 / 50
Vanishing polynomials of small degree
If the graph has some vertices of odd degree:
Theorem (A.V. 2016)
S3G vanishes (mod 2) on G precisely in one of the following two cases.
(A) G is a complete graph K4t+2, for some positive integer t, possiblyhaving 4u additional vertices and 4u(4t + 2) additional edges thatconnect these vertices to the above complete graph.(B) G is obtained by taking two complete graphs Ka, Kb and nadditional vertices satisfying one of the following conditions:
(B1) : a ≡ b ≡ 2 (mod 4) and n = 0;(B2) : a ≡ b ≡ n ≡ 1 (mod 4) except a = b = n = 1;(B3) : a ≡ b ≡ n ≡ 3 (mod 4);
every additional vertex must be fully connected to Ka and Kb.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 25 / 50
Vanishing polynomials of small degree
If the graph has some vertices of odd degree:
Theorem (A.V. 2016)
S3G vanishes (mod 2) on G precisely in one of the following two cases.
(A) G is a complete graph K4t+2, for some positive integer t, possiblyhaving 4u additional vertices and 4u(4t + 2) additional edges thatconnect these vertices to the above complete graph.(B) G is obtained by taking two complete graphs Ka, Kb and nadditional vertices satisfying one of the following conditions:
(B1) : a ≡ b ≡ 2 (mod 4) and n = 0;(B2) : a ≡ b ≡ n ≡ 1 (mod 4) except a = b = n = 1;(B3) : a ≡ b ≡ n ≡ 3 (mod 4);
every additional vertex must be fully connected to Ka and Kb.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 25 / 50
Vanishing polynomials of small degree
Old and new non-graceful graphs, as a consequence:
TheoremThe following graphs are non-graceful.Type (A) in the above theorem, provided t ≡ 2 (mod 4).Type (B1) with a + b ≡ 12 (mod 16); type (B2) with a + b ≡ 10 (mod16); type (B3) with a + b ≡ 14 (mod 16).
K7 K7
bbb
aaa��
������
a = b = 7 , n = 3
K5 K5
bbbbb
aaaaa����
����
a = b = n = 5
inglese italiano Traduci messaggio Disattiva per: inglese Ehrenfeucht-Fraïssé games on linear orders J K Truss (University of Leeds) This is joint work with Feresiano Mwesigye. In an n-move Ehrenfeucht-Fraïssé game on relational structures A and B, players I and II play alternately. On each move player I chooses an element of one of the structures, and player tried to ‘match it’ by choosing an element of the other structure. Player I does not have to choose from the same structure on each move. After n moves, points a1,a2, . . . ,an have been chosen in A, and points b1,b2, . . . ,bn have been chosen in B, and player II wins if the map taking ai to bi for each i is an isomorphism of induced substructures. Under these circumstances, we say that A and B are n-equivalent, written A ≡n B. The main facts about this situation are as follows. A and B are n-equivalent if and only if they satisfy the same sentences of quantifier depth at most n. Hence ≡n is an equivalence relation, and A and B are elementarily equivalent if and only if they are n-equivalent for all n. If the language is finite, then for each n there are just finitely many ≡n-classes. We consider linear orders, and also linear orders with colours, concentrating on estimates for the lengths of optimal representatives for finite coloured orders, optimal representatives for ordinals and certain special scattered liner orders.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 26 / 50
Vanishing polynomials of small degree
Degree 4
Not surprising::::::much
::::::more
::::::cases and
:::::many
:::::::::::subgraphs to check...
S4G ≡
∑p
H4px4
p +∑p,q
H31p,qx3
p xq +∑pq
H22pqx2
p x2q +
+∑p,qr
H211p,qr x
2p xqxr +
∑pqrs
H1111pqrs xpxqxr xs .
Strategy: (I) Find constraints for single vertices and pairs of vertices
(II) Constraints for subgraphs with 3 vertices – some ruled out by (I)
(III) Few subgraphs with 4 vertices survive. Last test: H1111pqrs ≡ 0
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 27 / 50
Vanishing polynomials of small degree
Degree 4
Not surprising::::::much
::::::more
::::::cases and
:::::many
:::::::::::subgraphs to check...
S4G ≡
∑p
H4px4
p +∑p,q
H31p,qx3
p xq +∑pq
H22pqx2
p x2q +
+∑p,qr
H211p,qr x
2p xqxr +
∑pqrs
H1111pqrs xpxqxr xs .
Strategy: (I) Find constraints for single vertices and pairs of vertices
(II) Constraints for subgraphs with 3 vertices – some ruled out by (I)
(III) Few subgraphs with 4 vertices survive. Last test: H1111pqrs ≡ 0
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 27 / 50
Vanishing polynomials of small degree
Degree 4
Not surprising::::::much
::::::more
::::::cases and
:::::many
:::::::::::subgraphs to check...
S4G ≡
∑p
H4px4
p +∑p,q
H31p,qx3
p xq +∑pq
H22pqx2
p x2q +
+∑p,qr
H211p,qr x
2p xqxr +
∑pqrs
H1111pqrs xpxqxr xs .
Strategy: (I) Find constraints for single vertices and pairs of vertices
(II) Constraints for subgraphs with 3 vertices – some ruled out by (I)
(III) Few subgraphs with 4 vertices survive. Last test: H1111pqrs ≡ 0
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 27 / 50
Vanishing polynomials of small degree
Degree 4
Not surprising::::::much
::::::more
::::::cases and
:::::many
:::::::::::subgraphs to check...
S4G ≡
∑p
H4px4
p +∑p,q
H31p,qx3
p xq +∑pq
H22pqx2
p x2q +
+∑p,qr
H211p,qr x
2p xqxr +
∑pqrs
H1111pqrs xpxqxr xs .
Strategy: (I) Find constraints for single vertices and pairs of vertices
(II) Constraints for subgraphs with 3 vertices – some ruled out by (I)
(III) Few subgraphs with 4 vertices survive. Last test: H1111pqrs ≡ 0
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 27 / 50
Vanishing polynomials of small degree
Degree 4
Not surprising::::::much
::::::more
::::::cases and
:::::many
:::::::::::subgraphs to check...
S4G ≡
∑p
H4px4
p +∑p,q
H31p,qx3
p xq +∑pq
H22pqx2
p x2q +
+∑p,qr
H211p,qr x
2p xqxr +
∑pqrs
H1111pqrs xpxqxr xs .
Strategy: (I) Find constraints for single vertices and pairs of vertices
(II) Constraints for subgraphs with 3 vertices – some ruled out by (I)
(III) Few subgraphs with 4 vertices survive. Last test: H1111pqrs ≡ 0
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 27 / 50
Vanishing polynomials of small degree
Some constraints for the degree 4
Constraint for single vertices: 0 ≤ δp ≤ 3 (mod 8)
Forbidden degrees of non-adjacent vertices:
(1,3), (2,2), (2,3), (3,3) (mod 4)
Forbidden degrees of adjacent vertices:
(0,0), (0,1), (3,3) (mod 4)
In particular, there is at most one vertex of degree 3 (mod 4)
etc. etc. (subgraphs with 3 vertices... nice to rephrase all this ascolour constraints)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 28 / 50
Vanishing polynomials of small degree
Some constraints for the degree 4
Constraint for single vertices: 0 ≤ δp ≤ 3 (mod 8)
Forbidden degrees of non-adjacent vertices:
(1,3), (2,2), (2,3), (3,3) (mod 4)
Forbidden degrees of adjacent vertices:
(0,0), (0,1), (3,3) (mod 4)
In particular, there is at most one vertex of degree 3 (mod 4)
etc. etc. (subgraphs with 3 vertices... nice to rephrase all this ascolour constraints)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 28 / 50
Vanishing polynomials of small degree
Some constraints for the degree 4
Constraint for single vertices: 0 ≤ δp ≤ 3 (mod 8)
Forbidden degrees of non-adjacent vertices:
(1,3), (2,2), (2,3), (3,3) (mod 4)
Forbidden degrees of adjacent vertices:
(0,0), (0,1), (3,3) (mod 4)
In particular, there is at most one vertex of degree 3 (mod 4)
etc. etc. (subgraphs with 3 vertices... nice to rephrase all this ascolour constraints)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 28 / 50
Vanishing polynomials of small degree
Some constraints for the degree 4
Constraint for single vertices: 0 ≤ δp ≤ 3 (mod 8)
Forbidden degrees of non-adjacent vertices:
(1,3), (2,2), (2,3), (3,3) (mod 4)
Forbidden degrees of adjacent vertices:
(0,0), (0,1), (3,3) (mod 4)
In particular, there is at most one vertex of degree 3 (mod 4)
etc. etc. (subgraphs with 3 vertices... nice to rephrase all this ascolour constraints)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 28 / 50
Vanishing polynomials of small degree
Some constraints for the degree 4
Constraint for single vertices: 0 ≤ δp ≤ 3 (mod 8)
Forbidden degrees of non-adjacent vertices:
(1,3), (2,2), (2,3), (3,3) (mod 4)
Forbidden degrees of adjacent vertices:
(0,0), (0,1), (3,3) (mod 4)
In particular, there is at most one vertex of degree 3 (mod 4)
etc. etc. (subgraphs with 3 vertices... nice to rephrase all this ascolour constraints)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 28 / 50
Vanishing polynomials of small degree
Degree 4: in the end...
Only two very small graphs have passed ALL the exams!
TheoremLet G be a graph having at least 4 edges and some odd vertices. Thepolynomial S4
G vanishes (mod 2) on G if and only if G is a 3-cycletogether with a further edge which can be either pendent or disjointfrom the cycle.
Here the advantages for gracefulness are NONE.
One graph is graceful, the other is not – exercise before sleeping.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 29 / 50
Vanishing polynomials of small degree
Degree 4: in the end...
Only two very small graphs have passed ALL the exams!
TheoremLet G be a graph having at least 4 edges and some odd vertices. Thepolynomial S4
G vanishes (mod 2) on G if and only if G is a 3-cycletogether with a further edge which can be either pendent or disjointfrom the cycle.
Here the advantages for gracefulness are NONE.
One graph is graceful, the other is not – exercise before sleeping.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 29 / 50
Vanishing polynomials of small degree
Degree 4: in the end...
Only two very small graphs have passed ALL the exams!
TheoremLet G be a graph having at least 4 edges and some odd vertices. Thepolynomial S4
G vanishes (mod 2) on G if and only if G is a 3-cycletogether with a further edge which can be either pendent or disjointfrom the cycle.
Here the advantages for gracefulness are NONE.
One graph is graceful, the other is not – exercise before sleeping.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 29 / 50
Vanishing polynomials of small degree
Degree 4: in the end...
Only two very small graphs have passed ALL the exams!
TheoremLet G be a graph having at least 4 edges and some odd vertices. Thepolynomial S4
G vanishes (mod 2) on G if and only if G is a 3-cycletogether with a further edge which can be either pendent or disjointfrom the cycle.
Here the advantages for gracefulness are NONE.
One graph is graceful, the other is not – exercise before sleeping.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 29 / 50
Final remarks on graceful polynomials
Final remarkson graceful polynomials
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 30 / 50
Final remarks on graceful polynomials
Final remarks
Degree 5 is being analysed...
Are there efficient techniques for higher degrees?
Why only vanishing (mod 2)?
We have a FAMILY OF POLYNOMIALS, let’s study it !
(modest aim: surpassing the fascinating popularity of thechromatic polynomial...)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 31 / 50
Final remarks on graceful polynomials
Final remarks
Degree 5 is being analysed...
Are there efficient techniques for higher degrees?
Why only vanishing (mod 2)?
We have a FAMILY OF POLYNOMIALS, let’s study it !
(modest aim: surpassing the fascinating popularity of thechromatic polynomial...)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 31 / 50
Final remarks on graceful polynomials
Final remarks
Degree 5 is being analysed...
Are there efficient techniques for higher degrees?
Why only vanishing (mod 2)?
We have a FAMILY OF POLYNOMIALS, let’s study it !
(modest aim: surpassing the fascinating popularity of thechromatic polynomial...)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 31 / 50
Final remarks on graceful polynomials
Final remarks
Degree 5 is being analysed...
Are there efficient techniques for higher degrees?
Why only vanishing (mod 2)?
We have a FAMILY OF POLYNOMIALS, let’s study it !
(modest aim: surpassing the fascinating popularity of thechromatic polynomial...)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 31 / 50
Final remarks on graceful polynomials
Final remarks
Degree 5 is being analysed...
Are there efficient techniques for higher degrees?
Why only vanishing (mod 2)?
We have a FAMILY OF POLYNOMIALS, let’s study it !
(modest aim: surpassing the fascinating popularity of thechromatic polynomial...)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 31 / 50
Final remarks on graceful polynomials
For Further Reading
J.A. GallianA Dynamic Survey of Graph LabelingElectr. J. Comb. 16, DS6 (2015).
A. RosaOn certain valuations of the vertices of a graphTheory of Graphs (Internat. Sympos. Rome, 1966) pp. 349-355,Gordon and Breach, New York; Dunod, Paris, 1967.
A. VietriNecessary conditions on graceful labels: a study case on treesand other examplesUtil. Math. 89 (2012), pp. 275-287.
A. VietriGraceful polynomials of graphs, and their vanishing (mod 2)March 2016, submitted.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 32 / 50
Real-graceful labellings
Real-graceful labellings
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 33 / 50
Real-graceful labellings
Fundamental ingredient in Rosa’s Theorem: double counting of labels(mod 2):
Is (mod 2) absolutely necessary?Counting (mod 2) drastically reduces the information, but...... it is the typical remedy agains the unpredictability of sign.
For, we can sum up (mod 2) the differences on the edges – and in themeantime sum up the labels on vertices – regardless of the SIGN.
Actually, there is at least an alternative way of managing theoscillations of sign!
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 34 / 50
Real-graceful labellings
Fundamental ingredient in Rosa’s Theorem: double counting of labels(mod 2):
Is (mod 2) absolutely necessary?Counting (mod 2) drastically reduces the information, but...
... it is the typical remedy agains the unpredictability of sign.
For, we can sum up (mod 2) the differences on the edges – and in themeantime sum up the labels on vertices – regardless of the SIGN.
Actually, there is at least an alternative way of managing theoscillations of sign!
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 34 / 50
Real-graceful labellings
Fundamental ingredient in Rosa’s Theorem: double counting of labels(mod 2):
Is (mod 2) absolutely necessary?Counting (mod 2) drastically reduces the information, but...... it is the typical remedy agains the unpredictability of sign.
For, we can sum up (mod 2) the differences on the edges – and in themeantime sum up the labels on vertices – regardless of the SIGN.
Actually, there is at least an alternative way of managing theoscillations of sign!
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 34 / 50
Real-graceful labellings
Fundamental ingredient in Rosa’s Theorem: double counting of labels(mod 2):
Is (mod 2) absolutely necessary?Counting (mod 2) drastically reduces the information, but...... it is the typical remedy agains the unpredictability of sign.
For, we can sum up (mod 2) the differences on the edges – and in themeantime sum up the labels on vertices – regardless of the SIGN.
Actually, there is at least an alternative way of managing theoscillations of sign!
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 34 / 50
Real-graceful labellings
Fundamental ingredient in Rosa’s Theorem: double counting of labels(mod 2):
Is (mod 2) absolutely necessary?Counting (mod 2) drastically reduces the information, but...... it is the typical remedy agains the unpredictability of sign.
For, we can sum up (mod 2) the differences on the edges – and in themeantime sum up the labels on vertices – regardless of the SIGN.
Actually, there is at least an alternative way of managing theoscillations of sign!
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 34 / 50
Real-graceful labellings
We can look at the SUM OF POWERS OF 2.
(however, any other base would work)
As an example, let us go back to the graceful graph K4:r rrr06 4
1@@��
Let us sum the terms 2a−b + 2b−a where {a,b} are the labels on eachedge – therefore, orientation does not count, similarly as counting(mod 2).
(20−4 + 24−0) + (24−1 + 21−4) + (26−1 + 21−6) + (20−6 + 26−0) +(20−1 + 21−0) + (24−6 + 26−4)...
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 35 / 50
Real-graceful labellings
We can look at the SUM OF POWERS OF 2.
(however, any other base would work)
As an example, let us go back to the graceful graph K4:r rrr06 4
1@@��
Let us sum the terms 2a−b + 2b−a where {a,b} are the labels on eachedge – therefore, orientation does not count, similarly as counting(mod 2).
(20−4 + 24−0) + (24−1 + 21−4) + (26−1 + 21−6) + (20−6 + 26−0) +(20−1 + 21−0) + (24−6 + 26−4)...
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 35 / 50
Real-graceful labellings
The contribution of all edges (because of gracefulness) is 2t + 2−t , forall t in 1,2, ...,e. (here e = 6).
TOTAL : 21 + 22 + · · ·+ 2e + 2−1 + 2−2 + · · ·+ 2−e =
=
(2e+1 − 1
2− 1− 20
)+
((2−1)e+1 − 1
2−1 − 1− (2−1)0
)=
= 2e+1 − 2−e − 1 .
In our example we obtain 27 − 2−6 − 1.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 36 / 50
Real-graceful labellings
The contribution of all edges (because of gracefulness) is 2t + 2−t , forall t in 1,2, ...,e. (here e = 6).
TOTAL : 21 + 22 + · · ·+ 2e + 2−1 + 2−2 + · · ·+ 2−e =
=
(2e+1 − 1
2− 1− 20
)+
((2−1)e+1 − 1
2−1 − 1− (2−1)0
)=
= 2e+1 − 2−e − 1 .
In our example we obtain 27 − 2−6 − 1.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 36 / 50
Real-graceful labellings
Theorem: (A.V. , 2011)Let G be a graph with e edges. If an injective labelling of the vertices ofG, with natural numbers not exceeding e, gives 2e+1 − 2−e − 1 as thetotal sum, then such a labelling is graceful.
Hint of proof via the Lemma:
Let n be a positive integer. Among all sums ofthe form∑
i ai2i equal ton and such that ai ∈ N for all i , the sum corresponding to the bynaryrepresentation of n minimises the quantity
∑i ai , and it is the only one
to have this property (induction on n etc.).Example:25 = 3 · 1 + 1 · 2 + 3 · 4 + 1 · 8 = 1 · 1 + 3 · 8 = 1 · 1 + 1 · 8 + 1 · 16 = ...
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 37 / 50
Real-graceful labellings
Theorem: (A.V. , 2011)Let G be a graph with e edges. If an injective labelling of the vertices ofG, with natural numbers not exceeding e, gives 2e+1 − 2−e − 1 as thetotal sum, then such a labelling is graceful.
Hint of proof via the Lemma:
Let n be a positive integer. Among all sums ofthe form∑
i ai2i equal ton and such that ai ∈ N for all i , the sum corresponding to the bynaryrepresentation of n minimises the quantity
∑i ai , and it is the only one
to have this property (induction on n etc.).Example:25 = 3 · 1 + 1 · 2 + 3 · 4 + 1 · 8 = 1 · 1 + 3 · 8 = 1 · 1 + 1 · 8 + 1 · 16 = ...
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 37 / 50
Real-graceful labellings
The Lemma has a Corollary which goes in our direction:
Corollary:Let P be a positive integer. If a non-trivial identity
∑1≤i≤P
2i +∑
1≤i≤P
(12
)i
=∑
1≤j≤J
aj2j +∑
1≤k≤K
a′k
(12
)k
holds for some positive integers J, K and with every aj , a′k in N, then∑j
aj +∑
k
a′k 6= 2P .
Now the proof of the theorem is closer...
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 38 / 50
Real-graceful labellings
The Lemma has a Corollary which goes in our direction:Corollary:Let P be a positive integer. If a non-trivial identity
∑1≤i≤P
2i +∑
1≤i≤P
(12
)i
=∑
1≤j≤J
aj2j +∑
1≤k≤K
a′k
(12
)k
holds for some positive integers J, K and with every aj , a′k in N, then∑j
aj +∑
k
a′k 6= 2P .
Now the proof of the theorem is closer...
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 38 / 50
Real-graceful labellings
Note 1: Almost an excercise... although it seems original!
Nota 2: Changing the base, we obtain other characterisations.
Let us now us particularly strange exponents: ...Definition: (MAIN)Let G be a graph with e edges. An injective mapping γ , from thevertices of G to the real interval [0,e] is a real-graceful labelling if∑
{u,v} spigolo di G
(2γ(u)−γ(v) + 2γ(v)−γ(u)
)= 2e+1 − 2−e − 1 .
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 39 / 50
Real-graceful labellings
Note 1: Almost an excercise... although it seems original!Nota 2: Changing the base, we obtain other characterisations.
Let us now us particularly strange exponents: ...Definition: (MAIN)Let G be a graph with e edges. An injective mapping γ , from thevertices of G to the real interval [0,e] is a real-graceful labelling if∑
{u,v} spigolo di G
(2γ(u)−γ(v) + 2γ(v)−γ(u)
)= 2e+1 − 2−e − 1 .
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 39 / 50
Real-graceful labellings
Note 1: Almost an excercise... although it seems original!Nota 2: Changing the base, we obtain other characterisations.
Let us now us particularly strange exponents: ...
Definition: (MAIN)Let G be a graph with e edges. An injective mapping γ , from thevertices of G to the real interval [0,e] is a real-graceful labelling if∑
{u,v} spigolo di G
(2γ(u)−γ(v) + 2γ(v)−γ(u)
)= 2e+1 − 2−e − 1 .
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 39 / 50
Real-graceful labellings
Note 1: Almost an excercise... although it seems original!Nota 2: Changing the base, we obtain other characterisations.
Let us now us particularly strange exponents: ...Definition: (MAIN)Let G be a graph with e edges. An injective mapping γ , from thevertices of G to the real interval [0,e] is a real-graceful labelling if∑
{u,v} spigolo di G
(2γ(u)−γ(v) + 2γ(v)−γ(u)
)= 2e+1 − 2−e − 1 .
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 39 / 50
Real-graceful labellings
Example: we know that C14 is not graceful;
let us nevertheless construct the oscillating classical labelling (alsowith the typical jump) but with an unknown:(14,0,13,1,12,2,11,3,10,5,9,6,8,x) .Now we calculate x by imposing that
2x−8 + 28−x + 2x−14 + 214−x = 21 + 2−1 + 26 + 2−6 ;
(we require that the differences arising from x contribute to the totalsum as the classical missing differences, ±1 e ±6)
(2x )2(1 + 26)− (2x ) · 4(
218 +52
212 + 26)
+ 16 · 218(26 + 1) = 0 ⇒
appr. x1 = 14.01 (NO because > 14), x2 = 7.99 (this is admissible!).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 40 / 50
Real-graceful labellings
Example: we know that C14 is not graceful;let us nevertheless construct the oscillating classical labelling (alsowith the typical jump) but with an unknown:(14,0,13,1,12,2,11,3,10,5,9,6,8,x) .
Now we calculate x by imposing that
2x−8 + 28−x + 2x−14 + 214−x = 21 + 2−1 + 26 + 2−6 ;
(we require that the differences arising from x contribute to the totalsum as the classical missing differences, ±1 e ±6)
(2x )2(1 + 26)− (2x ) · 4(
218 +52
212 + 26)
+ 16 · 218(26 + 1) = 0 ⇒
appr. x1 = 14.01 (NO because > 14), x2 = 7.99 (this is admissible!).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 40 / 50
Real-graceful labellings
Example: we know that C14 is not graceful;let us nevertheless construct the oscillating classical labelling (alsowith the typical jump) but with an unknown:(14,0,13,1,12,2,11,3,10,5,9,6,8,x) .Now we calculate x by imposing that
2x−8 + 28−x + 2x−14 + 214−x = 21 + 2−1 + 26 + 2−6 ;
(we require that the differences arising from x contribute to the totalsum as the classical missing differences, ±1 e ±6)
(2x )2(1 + 26)− (2x ) · 4(
218 +52
212 + 26)
+ 16 · 218(26 + 1) = 0 ⇒
appr. x1 = 14.01 (NO because > 14), x2 = 7.99 (this is admissible!).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 40 / 50
Real-graceful labellings
Example: we know that C14 is not graceful;let us nevertheless construct the oscillating classical labelling (alsowith the typical jump) but with an unknown:(14,0,13,1,12,2,11,3,10,5,9,6,8,x) .Now we calculate x by imposing that
2x−8 + 28−x + 2x−14 + 214−x = 21 + 2−1 + 26 + 2−6 ;
(we require that the differences arising from x contribute to the totalsum as the classical missing differences, ±1 e ±6)
(2x )2(1 + 26)− (2x ) · 4(
218 +52
212 + 26)
+ 16 · 218(26 + 1) = 0 ⇒
appr. x1 = 14.01 (NO because > 14), x2 = 7.99 (this is admissible!).
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 40 / 50
Real-graceful labellings
Note: When the cycle length increases, x tends to an adjacent label:the cycle behaves not that ORIGINALLY...But without the little jump, the equation would be not resolvable!
(14,0,13,1,12,2,11,3,10,4,9,5,8, x)
2x−8 + 28−x + 2x−14 + 214−x = 21 + 2−1 + 22 + 2−2 ⇒
(2x )2 · (1 + 26)− (2x ) · 27 · 212 + 16 · 224 + 16 · 218 = 0 ⇒
IMPOSSIBLE (∆ < 0) .Therefore, our expoinential equation makes some sense...
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 41 / 50
Real-graceful labellings
Note: When the cycle length increases, x tends to an adjacent label:the cycle behaves not that ORIGINALLY...But without the little jump, the equation would be not resolvable!
(14,0,13,1,12,2,11,3,10,4,9,5,8, x)
2x−8 + 28−x + 2x−14 + 214−x = 21 + 2−1 + 22 + 2−2 ⇒
(2x )2 · (1 + 26)− (2x ) · 27 · 212 + 16 · 224 + 16 · 218 = 0 ⇒
IMPOSSIBLE (∆ < 0) .
Therefore, our expoinential equation makes some sense...
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 41 / 50
Real-graceful labellings
Note: When the cycle length increases, x tends to an adjacent label:the cycle behaves not that ORIGINALLY...But without the little jump, the equation would be not resolvable!
(14,0,13,1,12,2,11,3,10,4,9,5,8, x)
2x−8 + 28−x + 2x−14 + 214−x = 21 + 2−1 + 22 + 2−2 ⇒
(2x )2 · (1 + 26)− (2x ) · 27 · 212 + 16 · 224 + 16 · 218 = 0 ⇒
IMPOSSIBLE (∆ < 0) .Therefore, our expoinential equation makes some sense...
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 41 / 50
Real-graceful labellings
Let us see what happens with complete graphs. Es. K5 .Let us leave only one label free: s s
sss
0
1
10
8
x
@@@ ���
BBBB
����
missing differences: ±3, ±4, ±5, ±6. Therefore:
2x−10 + 210−x + 2x−8 + 28−x + 2x−1 + 21−x + 2x + 2−x =∑
3≤i≤6
(2i + 2−i) .
Putting 2x = η , we have: 1541η2 − 123120η + 1313792 = 0. Appr.solutions for x : 3.66 e 6.07, both admissible.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 42 / 50
Real-graceful labellings
Let us see what happens with complete graphs. Es. K5 .Let us leave only one label free: s s
sss
0
1
10
8
x
@@@ ���
BBBB
����
missing differences: ±3, ±4, ±5, ±6. Therefore:
2x−10 + 210−x + 2x−8 + 28−x + 2x−1 + 21−x + 2x + 2−x =∑
3≤i≤6
(2i + 2−i) .
Putting 2x = η , we have: 1541η2 − 123120η + 1313792 = 0. Appr.solutions for x : 3.66 e 6.07, both admissible.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 42 / 50
Real-graceful labellings
Let us see what happens with complete graphs. Es. K5 .Let us leave only one label free: s s
sss
0
1
10
8
x
@@@ ���
BBBB
����
missing differences: ±3, ±4, ±5, ±6. Therefore:
2x−10 + 210−x + 2x−8 + 28−x + 2x−1 + 21−x + 2x + 2−x =∑
3≤i≤6
(2i + 2−i) .
Putting 2x = η , we have: 1541η2 − 123120η + 1313792 = 0. Appr.solutions for x : 3.66 e 6.07, both admissible.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 42 / 50
Real-graceful labellings
In order to label K6 we start for example with labels 15, 0, 1, 13, 4. Itremains to cover T = {5,6,7,8,10}. Let {pi} be the labels 15, 0, ... :
η2∑
1≤i≤5
215−pi − 215−10η(
210∑t∈T
2t +∑t∈T
210−t)
+∑
1≤i≤5
215+pi = 0 .
Appr. x = 5.81 e 10.87 (both admissible because smaller than 15).
For K7 we use 21, 0, 1, 19, 4, 14 and find x = 5.19 e 15.50.
Questions:1) Do there exist labels in some sense “ optimal ”?2) What is the largest number of integer labels one can ask for, whenthe order increases? (for K3, see later on)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 43 / 50
Real-graceful labellings
In order to label K6 we start for example with labels 15, 0, 1, 13, 4. Itremains to cover T = {5,6,7,8,10}. Let {pi} be the labels 15, 0, ... :
η2∑
1≤i≤5
215−pi − 215−10η(
210∑t∈T
2t +∑t∈T
210−t)
+∑
1≤i≤5
215+pi = 0 .
Appr. x = 5.81 e 10.87 (both admissible because smaller than 15).
For K7 we use 21, 0, 1, 19, 4, 14 and find x = 5.19 e 15.50.
Questions:1) Do there exist labels in some sense “ optimal ”?2) What is the largest number of integer labels one can ask for, whenthe order increases? (for K3, see later on)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 43 / 50
Real-graceful labellings
In order to label K6 we start for example with labels 15, 0, 1, 13, 4. Itremains to cover T = {5,6,7,8,10}. Let {pi} be the labels 15, 0, ... :
η2∑
1≤i≤5
215−pi − 215−10η(
210∑t∈T
2t +∑t∈T
210−t)
+∑
1≤i≤5
215+pi = 0 .
Appr. x = 5.81 e 10.87 (both admissible because smaller than 15).
For K7 we use 21, 0, 1, 19, 4, 14 and find x = 5.19 e 15.50.
Questions:1) Do there exist labels in some sense “ optimal ”?2) What is the largest number of integer labels one can ask for, whenthe order increases? (for K3, see later on)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 43 / 50
Real-graceful labellings
In order to label K6 we start for example with labels 15, 0, 1, 13, 4. Itremains to cover T = {5,6,7,8,10}. Let {pi} be the labels 15, 0, ... :
η2∑
1≤i≤5
215−pi − 215−10η(
210∑t∈T
2t +∑t∈T
210−t)
+∑
1≤i≤5
215+pi = 0 .
Appr. x = 5.81 e 10.87 (both admissible because smaller than 15).
For K7 we use 21, 0, 1, 19, 4, 14 and find x = 5.19 e 15.50.
Questions:1) Do there exist labels in some sense “ optimal ”?
2) What is the largest number of integer labels one can ask for, whenthe order increases? (for K3, see later on)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 43 / 50
Real-graceful labellings
In order to label K6 we start for example with labels 15, 0, 1, 13, 4. Itremains to cover T = {5,6,7,8,10}. Let {pi} be the labels 15, 0, ... :
η2∑
1≤i≤5
215−pi − 215−10η(
210∑t∈T
2t +∑t∈T
210−t)
+∑
1≤i≤5
215+pi = 0 .
Appr. x = 5.81 e 10.87 (both admissible because smaller than 15).
For K7 we use 21, 0, 1, 19, 4, 14 and find x = 5.19 e 15.50.
Questions:1) Do there exist labels in some sense “ optimal ”?2) What is the largest number of integer labels one can ask for, whenthe order increases? (for K3, see later on)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 43 / 50
Graceful hypersurfaces
Graceful hypersurfaces
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 44 / 50
Graceful hypersurfaces
Let us come back to the main definition.
Given a graph G with v vertices, all its real-graceful labellings (inparticular the CLASSICAL ones, if any) are, actually, points of ahypersurface of Rv depending on the graph.
Let us write 2xi−xj + 2xj−xi as
Xi
Xj+
Xj
Xi(Xs = 2xs ) .
Real-graceful labellings are the points X in the closed space [1,2e]v ,with pairwise distinct coordinates, belonging to the hypersurface
IG :∑
{ui ,uj} edge
(Xi
Xj+
Xj
Xi
)= 2e+1 − 2−e − 1 .
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 45 / 50
Graceful hypersurfaces
Let us come back to the main definition.
Given a graph G with v vertices, all its real-graceful labellings (inparticular the CLASSICAL ones, if any) are, actually, points of ahypersurface of Rv depending on the graph.Let us write 2xi−xj + 2xj−xi as
Xi
Xj+
Xj
Xi(Xs = 2xs ) .
Real-graceful labellings are the points X in the closed space [1,2e]v ,with pairwise distinct coordinates, belonging to the hypersurface
IG :∑
{ui ,uj} edge
(Xi
Xj+
Xj
Xi
)= 2e+1 − 2−e − 1 .
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 45 / 50
Graceful hypersurfaces
Let us come back to the main definition.
Given a graph G with v vertices, all its real-graceful labellings (inparticular the CLASSICAL ones, if any) are, actually, points of ahypersurface of Rv depending on the graph.Let us write 2xi−xj + 2xj−xi as
Xi
Xj+
Xj
Xi(Xs = 2xs ) .
Real-graceful labellings are the points X in the closed space [1,2e]v ,with pairwise distinct coordinates, belonging to the hypersurface
IG :∑
{ui ,uj} edge
(Xi
Xj+
Xj
Xi
)= 2e+1 − 2−e − 1 .
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 45 / 50
Graceful hypersurfaces
Classical graceful labellings correspond to the points of IG ∩ [1,2e]v
whose coordinates are distinct and in in {1,2,4,8, ...,2e}.
Problem A:Studying this particular class of hypersurfaces.Problem B:Are there non-graceful graphs as a consequence of the absence ofsuch points on IG? (very strong requirement...).Problem C:Can we reach a classical point along a curve which starts from anon-classical point, using a suitable recursion?Example: K4 . Start with 0, 0, 0, x . Find x (if any):
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 46 / 50
Graceful hypersurfaces
Classical graceful labellings correspond to the points of IG ∩ [1,2e]v
whose coordinates are distinct and in in {1,2,4,8, ...,2e}.Problem A:Studying this particular class of hypersurfaces.
Problem B:Are there non-graceful graphs as a consequence of the absence ofsuch points on IG? (very strong requirement...).Problem C:Can we reach a classical point along a curve which starts from anon-classical point, using a suitable recursion?Example: K4 . Start with 0, 0, 0, x . Find x (if any):
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 46 / 50
Graceful hypersurfaces
Classical graceful labellings correspond to the points of IG ∩ [1,2e]v
whose coordinates are distinct and in in {1,2,4,8, ...,2e}.Problem A:Studying this particular class of hypersurfaces.Problem B:Are there non-graceful graphs as a consequence of the absence ofsuch points on IG? (very strong requirement...).
Problem C:Can we reach a classical point along a curve which starts from anon-classical point, using a suitable recursion?Example: K4 . Start with 0, 0, 0, x . Find x (if any):
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 46 / 50
Graceful hypersurfaces
Classical graceful labellings correspond to the points of IG ∩ [1,2e]v
whose coordinates are distinct and in in {1,2,4,8, ...,2e}.Problem A:Studying this particular class of hypersurfaces.Problem B:Are there non-graceful graphs as a consequence of the absence ofsuch points on IG? (very strong requirement...).Problem C:Can we reach a classical point along a curve which starts from anon-classical point, using a suitable recursion?
Example: K4 . Start with 0, 0, 0, x . Find x (if any):
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 46 / 50
Graceful hypersurfaces
Classical graceful labellings correspond to the points of IG ∩ [1,2e]v
whose coordinates are distinct and in in {1,2,4,8, ...,2e}.Problem A:Studying this particular class of hypersurfaces.Problem B:Are there non-graceful graphs as a consequence of the absence ofsuch points on IG? (very strong requirement...).Problem C:Can we reach a classical point along a curve which starts from anon-classical point, using a suitable recursion?Example: K4 . Start with 0, 0, 0, x . Find x (if any):
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 46 / 50
Graceful hypersurfaces
Intersection between IG and the line η1 = η2 = η3 = 1, in R4 ; putη = 2x ;
3(1 + 1)2
1 · 1+ 3
(1 + η)2
1 · η= 27 − 2−6 − 1 + 2 · 6 ⇒
(A := 127− 164
... ) 3η2 + (6− A)η + 3 = 0
Appr. η = 40.30, η = 0.02. Only the first solutions falls in [1,64].Appr. x = 5.33.
r rrr00 0
5.33...@@�� ?
r rrr01 4
6@@��
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 47 / 50
Graceful hypersurfaces
Intersection between IG and the line η1 = η2 = η3 = 1, in R4 ; putη = 2x ;
3(1 + 1)2
1 · 1+ 3
(1 + η)2
1 · η= 27 − 2−6 − 1 + 2 · 6 ⇒
(A := 127− 164
... ) 3η2 + (6− A)η + 3 = 0
Appr. η = 40.30, η = 0.02. Only the first solutions falls in [1,64].Appr. x = 5.33.
r rrr00 0
5.33...@@�� ?
r rrr01 4
6@@��
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 47 / 50
Graceful hypersurfaces
Intersection between IG and the line η1 = η2 = η3 = 1, in R4 ; putη = 2x ;
3(1 + 1)2
1 · 1+ 3
(1 + η)2
1 · η= 27 − 2−6 − 1 + 2 · 6 ⇒
(A := 127− 164
... ) 3η2 + (6− A)η + 3 = 0
Appr. η = 40.30, η = 0.02. Only the first solutions falls in [1,64].Appr. x = 5.33.
r rrr00 0
5.33...@@�� ?
r rrr01 4
6@@��
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 47 / 50
Graceful hypersurfaces
Note: Starting with 4 equal coordinates leads to no solution.......................................................
the case of K3 (triangle, with labels x , y , z) is very intuitive; let us studythe corresponding surface:Put λ = 2x , µ = 2y , ν = 2z .
(λ+ µ)2
λµ+
(λ+ ν)2
λν+
(µ+ ν)2
µν=
1678
Homogeneous coordinates [λ, µ, ν] −→ projective curve in P2 .
Now ν 6= 0 ⇒ 8(X 2Y + XY 2 + X 2 + Y 2 + X + Y )− 119XY = 0
(X = λ/ν, Y = µ/ν). We obtain a real cubic Γ.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 48 / 50
Graceful hypersurfaces
Note: Starting with 4 equal coordinates leads to no solution.......................................................the case of K3 (triangle, with labels x , y , z) is very intuitive; let us studythe corresponding surface:
Put λ = 2x , µ = 2y , ν = 2z .
(λ+ µ)2
λµ+
(λ+ ν)2
λν+
(µ+ ν)2
µν=
1678
Homogeneous coordinates [λ, µ, ν] −→ projective curve in P2 .
Now ν 6= 0 ⇒ 8(X 2Y + XY 2 + X 2 + Y 2 + X + Y )− 119XY = 0
(X = λ/ν, Y = µ/ν). We obtain a real cubic Γ.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 48 / 50
Graceful hypersurfaces
Note: Starting with 4 equal coordinates leads to no solution.......................................................the case of K3 (triangle, with labels x , y , z) is very intuitive; let us studythe corresponding surface:Put λ = 2x , µ = 2y , ν = 2z .
(λ+ µ)2
λµ+
(λ+ ν)2
λν+
(µ+ ν)2
µν=
1678
Homogeneous coordinates [λ, µ, ν] −→ projective curve in P2 .
Now ν 6= 0 ⇒ 8(X 2Y + XY 2 + X 2 + Y 2 + X + Y )− 119XY = 0
(X = λ/ν, Y = µ/ν). We obtain a real cubic Γ.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 48 / 50
Graceful hypersurfaces
Note: Starting with 4 equal coordinates leads to no solution.......................................................the case of K3 (triangle, with labels x , y , z) is very intuitive; let us studythe corresponding surface:Put λ = 2x , µ = 2y , ν = 2z .
(λ+ µ)2
λµ+
(λ+ ν)2
λν+
(µ+ ν)2
µν=
1678
Homogeneous coordinates [λ, µ, ν] −→ projective curve in P2 .
Now ν 6= 0 ⇒ 8(X 2Y + XY 2 + X 2 + Y 2 + X + Y )− 119XY = 0
(X = λ/ν, Y = µ/ν). We obtain a real cubic Γ.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 48 / 50
Graceful hypersurfaces
Note: Starting with 4 equal coordinates leads to no solution.......................................................the case of K3 (triangle, with labels x , y , z) is very intuitive; let us studythe corresponding surface:Put λ = 2x , µ = 2y , ν = 2z .
(λ+ µ)2
λµ+
(λ+ ν)2
λν+
(µ+ ν)2
µν=
1678
Homogeneous coordinates [λ, µ, ν] −→ projective curve in P2 .
Now ν 6= 0 ⇒ 8(X 2Y + XY 2 + X 2 + Y 2 + X + Y )− 119XY = 0
(X = λ/ν, Y = µ/ν). We obtain a real cubic Γ.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 48 / 50
Graceful hypersurfaces
Es. (4,8) ∈ Γ. In fact log2 4, log2 8, log2 1 are graceful labels (note:ν = 1, so we have set a label equal to 0).Another “classical graceful” point: (2,8) (labels: 1,3,0 instead of 2,3,0).And so forth.
s s
ss
(4,8)
(8,4)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 49 / 50
Graceful hypersurfaces
Es. (4,8) ∈ Γ. In fact log2 4, log2 8, log2 1 are graceful labels (note:ν = 1, so we have set a label equal to 0).Another “classical graceful” point: (2,8) (labels: 1,3,0 instead of 2,3,0).And so forth.
s s
ss
(4,8)
(8,4)
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 49 / 50
Graceful hypersurfaces
For Further Reading
A. VietriReal-graceful labellings: a generalisation of graceful labellingsArs Comb. 102, 2011, pp. 359-364.
Andrea Vietri Some algebraic approaches to graceful labellings January, 9th, 2018 50 / 50