On the Gaussian Random Matrix Ensembles with Additional Symmetry Conditions
The Gaussian unitary random matrix ensembles satisfying some additional symmetry conditions are considered. The effect of these conditions on the limiting normalized counting measures and correlation functions is studied.
Збережено в:
| Дата: | 2006 |
|---|---|
| Автор: | |
| Формат: | Стаття |
| Мова: | English |
| Опубліковано: |
Інститут математики НАН України
2006
|
| Назва видання: | Symmetry, Integrability and Geometry: Methods and Applications |
| Онлайн доступ: | http://dspace.nbuv.gov.ua/handle/123456789/146451 |
| Теги: |
Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Цитувати: | On the Gaussian Random Matrix Ensembles with Additional Symmetry Conditions / V. Vasilchuk // Symmetry, Integrability and Geometry: Methods and Applications. — 2006. — Т. 2. — Бібліогр.: 9 назв. — англ. |
Репозитарії
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
irk-123456789-146451 |
|---|---|
| record_format |
dspace |
| spelling |
irk-123456789-1464512019-02-10T01:25:21Z On the Gaussian Random Matrix Ensembles with Additional Symmetry Conditions Vasilchuk, V. The Gaussian unitary random matrix ensembles satisfying some additional symmetry conditions are considered. The effect of these conditions on the limiting normalized counting measures and correlation functions is studied. 2006 Article On the Gaussian Random Matrix Ensembles with Additional Symmetry Conditions / V. Vasilchuk // Symmetry, Integrability and Geometry: Methods and Applications. — 2006. — Т. 2. — Бібліогр.: 9 назв. — англ. 1815-0659 2000 Mathematics Subject Classification: 15A52; 60A10; 82B41 http://dspace.nbuv.gov.ua/handle/123456789/146451 en Symmetry, Integrability and Geometry: Methods and Applications Інститут математики НАН України |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| language |
English |
| description |
The Gaussian unitary random matrix ensembles satisfying some additional symmetry conditions are considered. The effect of these conditions on the limiting normalized counting measures and correlation functions is studied. |
| format |
Article |
| author |
Vasilchuk, V. |
| spellingShingle |
Vasilchuk, V. On the Gaussian Random Matrix Ensembles with Additional Symmetry Conditions Symmetry, Integrability and Geometry: Methods and Applications |
| author_facet |
Vasilchuk, V. |
| author_sort |
Vasilchuk, V. |
| title |
On the Gaussian Random Matrix Ensembles with Additional Symmetry Conditions |
| title_short |
On the Gaussian Random Matrix Ensembles with Additional Symmetry Conditions |
| title_full |
On the Gaussian Random Matrix Ensembles with Additional Symmetry Conditions |
| title_fullStr |
On the Gaussian Random Matrix Ensembles with Additional Symmetry Conditions |
| title_full_unstemmed |
On the Gaussian Random Matrix Ensembles with Additional Symmetry Conditions |
| title_sort |
on the gaussian random matrix ensembles with additional symmetry conditions |
| publisher |
Інститут математики НАН України |
| publishDate |
2006 |
| url |
http://dspace.nbuv.gov.ua/handle/123456789/146451 |
| citation_txt |
On the Gaussian Random Matrix Ensembles with Additional Symmetry Conditions / V. Vasilchuk // Symmetry, Integrability and Geometry: Methods and Applications. — 2006. — Т. 2. — Бібліогр.: 9 назв. — англ. |
| series |
Symmetry, Integrability and Geometry: Methods and Applications |
| work_keys_str_mv |
AT vasilchukv onthegaussianrandommatrixensembleswithadditionalsymmetryconditions |
| first_indexed |
2025-07-11T00:00:10Z |
| last_indexed |
2025-07-11T00:00:10Z |
| _version_ |
1837306500698079232 |
| fulltext |
Symmetry, Integrability and Geometry: Methods and Applications Vol. 2 (2006), Paper 007, 12 pages
On the Gaussian Random Matrix Ensembles
with Additional Symmetry Conditions
Vladimir VASILCHUK
B. Verkin Institute for Low Temperature Physics and Engineering,
47 Lenin Ave., Kharkiv, 61103 Ukraine
E-mail: vasilchuk@ilt.kharkov.ua
Received October 31, 2005, in final form January 06, 2006; Published online January 21, 2006
Original article is available at http://www.emis.de/journals/SIGMA/2006/Paper007/
Abstract. The Gaussian unitary random matrix ensembles satisfying some additional sym-
metry conditions are considered. The effect of these conditions on the limiting normalized
counting measures and correlation functions is studied.
Key words: random matrices; Gaussian unitary ensemble
2000 Mathematics Subject Classification: 15A52; 60A10; 82B41
1 Introduction and main results
Let us consider a standard 2n × 2n Gaussian Unitary Ensemble (GUE) of Hermitian random
matrices Wn:
Wn = W †
n, (Wn)xy =
1√
2n
(ξxy + iηxy) , (1)
where ξxy, ηxy, x, y = −n, . . . ,−1, 1, . . . , n are i.i.d. Gaussian random variables with zero mean
and variance 1/2. Consider also the normalized eigenvalue counting measure (NCM) Nn of the
ensemble (1), defined for any Borel set ∆ ⊂ R by the formula
Nn(∆) =
#{λi ∈ ∆}
2n
, (2)
where λi, i = 1, . . . , 2n are the eigenvalues of Wn.
Suppose now that ensemble (1) has also an additional symmetry of negative (positive) indices
x and y. We consider four different cases of symmetry:
1. (Wn)xy = (Wn)−y−x , (3)
2. (Wn)xy = (Wn)−x−y , (4)
3. (Wn)xy = (Wn)−xy , (5)
4. (Wn)xy = (Wn)y−x . (6)
The Gaussian unitary ensemble and Gaussian orthogonal ensemble (GOE) was considered
in numerous papers (see e.g. [4]). The Gaussian unitary ensemble with additional symmetry of
type (3) was proposed in the papers [1, 3] as an approach to the weak disorder regime in the
Anderson model. This ensemble was also considered in the papers [2, 9]. In all these papers
ensemble (3) was called as flip matrix model and studied by some supersymmetry approach and
moments method. In this paper an approach is proposed that is simpler and the same for all
mailto:vasilchuk@ilt.kharkov.ua
http://www.emis.de/journals/SIGMA/2006/Paper007/
2 V. Vasilchuk
four cases (3)–(6). This approach is a version of technique initially proposed in [7] and developed
in the papers [5, 4, 6, 8].
Using this technique we obtain the following results.
First two ensembles (3) and (4) are GOE-like.
Proposition 1. The NCMs N
(1)
n and N
(2)
n of the ensembles (3) and (4) converge weakly with
probability 1 to the semi-circle law Nsc
Nsc(dλ) = (2π)−1
√
4− λ2χ[−2,2](λ)dλ
and the n−1-asymptotics of the correlation functions
F (i)
n (z1, z2) = E
{(
g(i)
n (z1)− Eg(i)
n (z1)
)(
g(i)
n (z2)− Eg(i)
n (z2)
)}
, i = 1, 2
of their Stieltjes transforms
g(i)
n (z) =
∫ ∞
−∞
N
(i)
n (dλ)
λ− z
, Im z > 0, i = 1, 2
coincide with corresponding 2n× 2n-GOE asymptotic (2n)−2 S(z1, z2) [4]:
F (i)
n (z1, z2) = (2n)−2 S(z1, z2) + o
(
n−2
)
,
S(z1, z2) =
2
(1− f2
sc(z1)) (1− f2
sc(z2))
(
fsc(z1)− fsc(z2)
z1 − z2
)2
, (7)
where
fsc(z) =
∫ ∞
−∞
Nsc(dλ)
λ− z
, Im z > 0
is the Stieltjes transform of the semi-circle law Nsc.
The fourth ensemble (6) is GUE-like:
Proposition 2. The NCM N
(4)
n of the ensemble (6) converges weakly with probability 1 to the
semi-circle law Nsc and the n−1-asymptotic of the correlation function of its Stieltjes transform
coincides with (7) divided by 2 (i.e. GUE asymptotic).
As for the third ensemble, the additional symmetry produces new limiting NCM and corre-
lation function:
Theorem 1. The NCM N
(3)
n of the ensemble (5) converges weakly with probability 1 to the
limiting non-random measure N
N(dλ) =
1
4
δ(λ)dλ +
1
4π
√
6− (λ2 − λ−2)χ[−λ+,−λ−]∪[λ−,λ+](λ)dλ, (8)
where λ± =
√
3± 2
√
2 and the n−1-asymptotic of the correlation function of its Stieltjes trans-
form is given by the formula
F (3)
n (z1, z2) = (2n)−2 C(z1, z2) + o
(
n−2
)
,
C(z1, z2) =
(
2
f2(z1) + f2(z2)
f(z1)f(z2)(z1 − z2)2
+
z2f(z2) + z1f(z1)
2z2
1z
2
2f(z1)f(z2)
)∏
k=1,2
(
zk + z−1
k + 4f(zk)
)−1
, (9)
where f(z) is Stieltjes transform of the limiting measure N .
This result is somewhat unexpected for the Hermitian Gaussian random matrix ensemble
with the rather large number (of the order n2) of independent random parameters. But it shows
how much the additional symmetry may affect the asymptotic behavior of the eigenvalues.
On the GUEs with Additional Symmetries 3
2 The limiting NCMs
In this section we consider the limiting normalized countable measures of the ensembles (3)–(6).
In that follows we use the notations
G(z) = (Wn − z)−1 , ĝ(z) =
1
2n
n∑
j=−n
Gj−j(z),
g(z) =
1
2n
TrG(z) =
n∑
j=−n
Gjj(z),
and 〈·〉 to denote the average over GUE. We also use the resolvent identity
G(z) = −z−1I + z−1WnG(z)
and the Novikov–Furutsu formula for the complex Gaussian random variable ζ = ξ + iη with
zero mean and variance 1, and for the continuously differentiable function q(x, x)
Eζq(ζ, ζ) = E
∂
∂ζ
q(ζ, ζ), (10)
where ∂
∂ζ
= 1
2
(
∂
∂ξ + i ∂
∂η
)
.
We will perform our calculations in parallel for all four ensembles. First, let us observe that
properties (3)–(5) are valid not only for the matrices of ensembles (3)–(5) but for their powers
and hence also for their resolvents. Indeed, using induction by m and the symmetry of summing
index we obtain:
1.
(
Wm+1
n
)
jk
=
n∑
l=−n
(Wn)jl
(
Wm+1
n
)
lk
=
n∑
l=−n
(Wn)j−l
(
Wm+1
n
)
−lk
=
n∑
l=−n
(Wn)l−j
(
Wm+1
n
)
−kl
=
(
Wm+1
n
)
−k−j
,
Thus, Gjk(z) = G−k−j(z).
2.
(
Wm+1
n
)
jk
=
n∑
l=−n
(Wn)jl
(
Wm+1
n
)
lk
=
n∑
l=−n
(Wn)j−l
(
Wm+1
n
)
−lk
=
n∑
l=−n
(Wn)−jl
(
Wm+1
n
)
l−k
=
(
Wm+1
n
)
−j−k
,
Thus, Gjk(z) = G−j−k(z).
3.
(
Wm+1
n
)
jk
=
n∑
l=−n
(Wn)jl
(
Wm+1
n
)
lk
=
n∑
l=−n
(Wn)−jl
(
Wm+1
n
)
lk
=
(
Wm+1
n
)
−jk
.
Thus,
Gjk(z) = G−jk(z)− z−1(δjk − δ−jk) (11)
4 V. Vasilchuk
and, hence,
g(z) = ĝ(z)− z−1, where ĝ(z) =
n∑
r=−n
Gr−r(z).
Unfortunately, there is no any such property for the fourth ensemble.
Now using the resolvent identity for the average 〈Gpq(z)〉, relation (10) and formula for the
derivative of the resolvent
G′(z) ·X = −G(z)XG(z), (12)
we obtain
〈Gpq(z)〉 = −z−1δpq + z−1
〈
(WnG(z))pq
〉
= −z−1δpq + z−1 1√
2n
n∑
r=−n
〈
1
2
(
∂
∂ξpr
+ i
∂
∂ηpr
)
Grq(z)
〉
= −z−1δpq + z−1 1√
2n
n∑
r,j,k=−n
〈
Grj(z)
(
W ′
n
)
jk
Gkq(z)
〉
,
where W ′
n = 1
2
(
∂
∂ξpr
+ i ∂
∂ηpr
)
Wn. Now we calculate W ′
n for all four ensembles:
1. W ′
n =
1
2
√
2n
(
δjpδkr + δjrδkp + δj−rδk−p + δj−pδk−r
−δjpδkr + δjrδkp − δj−rδk−p + δj−pδk−r
)
=
1√
2n
(δjrδkp + δj−pδk−r) ;
2. W ′
n =
1
2
√
2n
(
δjpδkr + δjrδkp + δj−pδk−r + δj−rδk−p
−δjpδkr + δjrδkp − δj−pδk−r + δj−rδk−p
)
=
1√
2n
(δjrδkp + δj−rδk−p) ;
3. W ′
n =
1
2
√
2n
(
δjpδkr + δjrδkp + δj−pδkr + δjrδk−p
−δjpδkr + δjrδkp − δj−pδkr + δjrδk−p
)
=
1√
2n
(δjrδkp + δjrδk−p) ;
4. W ′
n =
1
2
√
2n
(
δjpδkr + δjrδkp + δjrδk−p + δj−pδkr
−δjpδkr + δjrδkp − δjrδk−p + δj−pδkr
)
=
1√
2n
(δjrδkp + δj−pδkr) .
Using these formulas, we obtain the following relations:
1. 〈Gpq(z)〉 = −z−1δpq − z−1 〈g(z)Gpq(z)〉 − z−1
〈
1
2n
n∑
r=−n
Gr−p(z)G−rq(z)
〉
;
2. 〈Gpq(z)〉 = −z−1δpq − z−1 〈g(z)Gpq(z)〉 − z−1 〈ĝ(z)G−pq(z)〉 ;
3. 〈Gpq(z)〉 = −z−1δpq − z−1 〈g(z)Gpq(z)〉 − z−1 〈g(z)G−pq(z)〉 ;
4. 〈Gpq(z)〉 = −z−1δpq − z−1 〈g(z)Gpq(z)〉 − z−1
〈
1
2n
n∑
r=−n
Gr−p(z)Grq(z)
〉
. (13)
On the GUEs with Additional Symmetries 5
Now we put p = q in all four cases and p = −q another time in the second case, and apply
1
2n
n∑
p=−n
. Thus, using also the additional symmetries of the resolvents of ensembles (3)–(5), we
obtain:
1. 〈g(z)〉 = −z−1
(
1 + 〈g(z)〉2
)
− z−1
[
1
2n
〈
1
2n
TrG2(z)
〉
+ 〈g◦(z)g(z)〉
]
,
where g◦(z) = g(z)− 〈g(z)〉;
2. 〈g(z)〉 = −z−1
(
1 + 〈g(z)〉2 +
〈
ĝ2(z)
〉)
− z−1 [〈g◦(z)g(z)〉+ 〈ĝ◦(z)ĝ(z)〉] ,
〈ĝ(z)〉 = −2z−1 〈g(z)〉 〈ĝ(z)〉 − 2z−1 〈g◦(z)ĝ(z)〉 ;
3. 〈g(z)〉 = −z−1
(
1 + 〈g(z)〉2 + 〈g(z)〉 〈ĝ(z)〉
)
− z−1 [〈g◦(z)g(z)〉+ 〈ĝ◦(z)ĝ(z)〉] ,
ĝ(z) = g(z) + z−1; (14)
4. 〈g(z)〉 = −z−1
(
1 + 〈g(z)〉2
)
− z−1
[
1
2n
〈
1
2n
TrP (z)G(z)
〉
+ 〈g◦(z)g(z)〉
]
,
where matrix P (z) is defined by Pxy(z) = Gy−x(z).
In the appendix we prove that the variances of random variables g(z) in all cases above are of
the order O(n−2) uniformly in z for some compact in C± (as well as the variance of ĝ(z) in the
second case). Besides, using Schwartz inequality for the matrix scalar product (A,B) = TrAB,
we obtain∣∣∣∣ 1
2n
TrP (z)G(z)
∣∣∣∣ ≤ ( 1
2n
TrP (z)P †(z)
)1/2( 1
2n
TrG(z)G†(z)
)1/2
≤ 1
|Im z|2
,∣∣∣∣ 1
2n
TrG2(z)
∣∣∣∣ ≤ 1
|Im z|2
.
Thus, all terms in square brackets in all four cases are at least of the order O(n−1). Hence, in
the first and in the fourth cases we obtain the following limiting equation:
f(z) = −z−1
(
1 + f2(z)
)
, (15)
which is the equation for fsc(z) — the Stieltjes transform of the semi-circle Law.
Besides, since
g(z) = −z−1
pq − z−1 1
2n
Tr (WnG(z)) ,∣∣∣∣ 1
2n
Tr (WnG(z))
∣∣∣∣ ≤ 1
|Im z|
(
1
2n
TrWnW †
n
)1/2
,〈(
1
2n
TrWnW †
n
)1/2
〉
≤
〈
1
2n
TrWnW †
n
〉1/2
≤ 1,
then for all z with e.g. |Im z| ≥ 3 uniformly in n we have in all cases∣∣1 + 2z−1 〈g(z)〉
∣∣ > 1
2
. (16)
Thus, in the second case 〈ĝ(z)〉 of the order O(n−2):
〈ĝ(z)〉 = −2z−1
(
1 + 2z−1 〈g(z)〉
)−1 〈g◦(z)ĝ(z)〉 .
Hence, the second case lead to the same limiting equation (15).
6 V. Vasilchuk
As for the third case, it leads to the following equation
f(z) = −z−1
(
1 + 2f2(z) + z−1f(z)
)
. (17)
Its solution in the class of Nevanlinna functions is the Stieltjes transform of the measure (8).
The convergence with probability one in all four cases follows from the bounds for the vari-
ances in the section bellow and the Borel–Cantelli lemma.
3 The correlation functions
As in the previous section, we perform our calculations in parallel for all four ensembles.
Using the resolvent identity for the average 〈g◦(z1)Gpq(z2)〉, relations (10) and (12), we obtain
〈g◦(z1)Gpq(z2)〉 = z−1
2
1√
2n
n∑
r,j,k=−n
〈
g◦(z1)Grj(z2)
(
W ′
n
)
jk
Gkq(z2)
〉
+ z−1
2
1
(2n)3/2
n∑
l.r,j,k=−n
〈
Glj(z1)
(
W ′
n
)
jk
Gkl(z1)Grq(z2)
〉
.
Substituting in this relation the value of W ′
n in all four cases and using the symmetries of the
resolvents, we obtain
1. 〈g◦(z1)Gpq(z2)〉 = −z−1
2 〈g◦(z1)g(z2)Gpq(z2)〉 − z−1
2
〈
g◦(z1)
1
2n
G2
pq(z2)
〉
− z−1
2
1
(2n)2
(〈(
G2(z1)G(z2)
)
pq
〉
+
〈(
G(z2)G2(z1)
)
−q−p
〉)
;
2. 〈g◦(z1)Gpq(z2)〉 = −z−1
2 〈g◦(z1)g(z2)Gpq(z2)〉 − z−1
2 〈g◦(z1)ĝ(z2)G−pq(z2)〉
− z−1
2
1
(2n)2
(〈(
G2(z1)G(z2)
)
pq
〉
+
〈(
G(z2)G2(z1)
)
−p−q
〉)
;
3. 〈g◦(z1)Gpq(z2)〉 = −z−1
2 〈g◦(z1)g(z2)Gpq(z2)〉 − z−1
2 〈g◦(z1)g(z2)G−pq(z2)〉
− z−1
2
1
(2n)2
(〈(
G2(z1)G(z2)
)
pq
〉
+
〈(
G2(z1)G(z2)
)
−pq
〉)
;
4. 〈g◦(z1)Gpq(z2)〉 = −z−1
2 〈g◦(z1)g(z2)Gpq(z2)〉 − z−1
2
〈
g◦(z1)
1
2n
n∑
r=−n
Gr−p(z2)Grq(z2)
〉
− z−1
2
1
(2n)2
(〈(
G2(z1)G(z2)
)
pq
〉
+
〈
n∑
r=−n
(
G2(z1)
)
r−p
Grq(z2)
〉)
.
Then we put p = q in all four cases and p = −q another time in the second case, and apply
1
2n
n∑
p=−n
and obtain
1. 〈g◦(z1)g(z2)〉 = −2z−1
2 〈g(z2)〉 〈g◦(z1)g(z2)〉
− z−1
2
2
(2n)2
〈
1
2n
TrG2(z1)G(z2)
〉
+ r1,n, (18)
where
r1,n = −z−1
2
(〈
g◦(z1) (g◦(z2))
2
〉
+
1
2n
〈
g◦(z1)
1
2n
TrG2(z2)
〉)
; (19)
On the GUEs with Additional Symmetries 7
2. 〈g◦(z1)g(z2)〉 = −2z−1
2 (〈g(z2)〉 〈g◦(z1)g(z2)〉+ 〈ĝ(z2)〉 〈g◦(z1)ĝ(z2)〉)
− z−1
2
2
(2n)2
〈
1
2n
TrG2(z1)G(z2)
〉
+ r2,n,
〈ĝ◦(z1)ĝ(z2)〉 = −2z−1
2 (〈g(z2)〉 〈ĝ◦(z1)ĝ(z2)〉+ 〈ĝ(z2)〉 〈g◦(z1)ĝ(z2)〉)
− z−1
2
2
(2n)2
〈
1
2n
TrG2(z1)G(z2)
〉
+ r3,n,
where
r2,n = −z−1
2
(〈
g◦(z1) (g◦(z2))
2
〉
+
〈
g◦(z1) (ĝ◦(z2))
2
〉)
,
r3,n = −2z−1
2 〈ĝ◦(z1)ĝ◦(z2)g◦(z2)〉 ;
3. 〈g◦(z1)g(z2)〉 = −4z−1
2 〈g(z2)〉 〈g◦(z1)g(z2)〉 − z−2
2 〈g◦(z1)g(z2)〉
− z−1
2
1
(2n)2
〈
1
2n
TrG2(z1)G(z2)
〉
− z−1
2
1
(2n)2
〈
1
2n
n∑
p=−n
(
G2(z1)G(z2)
)
−pp
〉
+ r4,n,
where
r4,n = −2z−1
2
〈
g◦(z1) (g◦(z2))
2
〉
;
4. 〈g◦(z1)g(z2)〉 = −2z−1
2 〈g(z2)〉 〈g◦(z1)g(z2)〉
− z−1
2
1
(2n)2
〈
1
2n
TrG2(z1)G(z2)
〉
+ r5,n,
where
r5,n = −z−1
2
1
2n
〈
g◦(z1)
1
2n
TrP (z2)G(z2)
〉
− z−1
2
〈
g◦(z1) (g◦(z2))
2
〉
− z−1
2
1
(2n)2
〈
1
2n
n∑
r,p=−n
(
G2(z1)
)
r−p
Grp(z2)
〉
. (20)
As we show in the appendix, all rj,n, j = 1, . . . , 5 are of the order o(n−2). Thus, as one can
easily show, all correlation functions F (z1, z2) = 〈g◦(z1)g(z2)〉 above are of the order O(n−2).
Moreover, since 〈ĝ(z2)〉 is of the order O(n−2) in the second case, its easy to see that cases one
and two lead to the same relation for F (z1, z2)
F (z1, z2) = −2z−1
2 〈g(z2)〉F (z1, z2)− z−1
2
2
(2n)2
〈
1
2n
TrG2(z1)G(z2)
〉
+ o
(
n−2
)
. (21)
As to the case four, it leads to
F (z1, z2) = −2z−1
2 〈g(z2)〉F (z1, z2)− z−1
2
1
(2n)2
〈
1
2n
TrG2(z1)G(z2)
〉
+ o
(
n−2
)
. (22)
Besides, due to the resolvent identity we have
1
2n
TrG2(z1)G(z2) =
1
z1 − z2
(
1
2n
TrG2(z1)−
g(z1)− g(z2)
z1 − z2
)
. (23)
8 V. Vasilchuk
In addition, as we show in the appendix, in these cases〈
1
2n
TrG2(z)
〉
=
〈g(z)〉
1− 〈g(z)〉2
+ O
(
n−1
)
. (24)
Thus, substituting in the relations (21), (22) the expressions (23), (24) and using the equa-
tion (15) for the limit of 〈g(z)〉, we obtain in the cases one and two the GOE correlator asymp-
totic (7) and in the case four the twice less GUE asymptotic.
To treat the third case we use (11) and obtain that
1
2n
n∑
p=−n
(
G2(z1)G(z2)
)
−pp
=
1
2n
TrG2(z1)G(z2) +
1
z2
1z2
.
This gives the following relation for F (z1, z2)
F (z1, z2) =
(
−4
〈g(z2)〉
z2
− z−2
2
)
F (z1, z2)
− z−1
2
(2n)2
(
1
z2
1z2
+
1
z1 − z2
〈
1
2n
TrG2(z1)−
g(z1)− g(z2)
z1 − z2
〉)
+ o
(
n−2
)
. (25)
We show also in the appendix that in this case〈
1
2n
TrG2(z)
〉
= −〈g(z)〉
z
1− z−2
1 + 〈g(z)〉 z−1 + z−2
+ o
(
n−2
)
.
Substituting this relation in (25) we obtain
F (z1, z2) =
1
n2
− 1
(z1z2)2
+ 2
z1
f(z1)−f(z2)
(z1−z2)2
1 + z−2
1 + 4z−1
1 f(z1)
− 2
1− z−2
2
z1 − z2
f(z2)
z1z2
∏
k=1,2
(
1 + z−2
k + 4
f(zk)
zk
)+ o
(
n−2
)
.
Then, using the equation (17), we rewrite this relation in the form (9).
4 Conclusion
The purpose of this paper was to answer the question: “Can the additional symmetry prop-
erties influence on the asymptotic behavior of eigenvalue distribution of GUE?” The negative
answer for the three cases of additional symmetry is not surprising, as these symmetries leave
the number of independent random parameters of the order n2. The effect when in one case the
additional symmetry essentially changes the limiting eigenvalue counting measure is very unex-
pected, especially the appearance of the gap in the support of limiting NCM. Unfortunately, the
physical application of this effect is unknown to the author, though one of the other considered
ensembles (flip matrix model) was used as an approach to weak coupling regime of the Anderson
model.
A Appendix
Proposition 3. The variance v =
〈
|g◦(z)|2
〉
is of the order O(n−2) in all four cases, and the
terms rj,n, j = 1, . . . , 5 are of the order o(n−2).
On the GUEs with Additional Symmetries 9
Proof. First we proof that the variance is of the order O(n−2) in all four cases. Indeed, in the
first case, using (18) with z2 = z1 = z, we obtain
v(1 + 2z−1 〈g(z)〉) = −z−1
2
2
(2n)2
〈
1
2n
TrG2(z1)G(z2)
〉
+ r1,n.
Besides, using the Schwartz inequality we obtain from (19)
r1,n ≤ |z|−1
(
1
|Im z|
v +
1
2n |Im z|2
v1/2
)
.
Thus, due to the bounds (16) and∣∣∣∣ 1
2n
TrG2(z1)G(z2)
∣∣∣∣ ≤ 1
|Im z|3
, (26)
we have for |Im z| ≥ 3 the inequality
v ≤ 2
9 (2n)2
+
1
2n
v1/2,
which leads to v = O(n−2). For the other cases the proofs are analogous.
To prove r1,n = o(n−2) for
r1,n = −z−1
2
(〈
g◦(z1) (g◦(z2))
2
〉
+
1
2n
〈
g◦(z1)
1
2n
TrG2(z2)
〉)
,
we rewrite the second term in the parentheses as
1
2n
〈
g◦(z1)
1
2n
TrG2(z2)
〉
=
1
2n
〈
g◦(z1)
1
2n
Tr
∂
∂z2
G(z2)
〉
=
1
2n
∂
∂z2
〈g◦(z1)g◦(z2)〉 .
Since the value 〈g◦(z1)g◦(z2)〉 is analytical and uniformly in n bounded for |Im z1,2| ≥ 3, and
since, due to the Schwartz inequality |〈g◦(z1)g◦(z2)〉| ≤ v = O(n−2), its derivative on z2 is also
of the order O(n−2) and hence the second term is of the order O(n−3).
To prove that the first term is o(n−2) let us consider〈
|g◦(z)|4
〉
=
〈
(g◦(z1)g◦(z2))
2
〉
= 〈R◦g(z2)〉 , R ≡ (g◦(z1))
2 g◦(z2), z1 = z2 = z.
Then, using the resolvent identity for the average 〈R◦Gpq(z2)〉, relations (10) and (12), we obtain
〈R◦Gpq(z2)〉 = z−1
2
1√
2n
n∑
r,j,k=−n
〈
R◦Grj(z2)
(
W ′
n
)
jk
Gkq(z2)
〉
+ z−1
2
2
(2n)3/2
n∑
l.r,j,k=−n
〈
g◦(z1)g◦(z2)Glj(z1)
(
W ′
n
)
jk
Gkl(z1)Grq(z2)
〉
+ z−1
2
1
(2n)3/2
n∑
l.r,j,k=−n
〈
(g◦(z1))
2 Glj(z2)
(
W ′
n
)
jk
Gkl(z2)Grq(z2)
〉
.
Substituting in this relation the value of W ′
n and using the symmetry of the resolvent we obtain
〈R◦Gpq(z2)〉 = −z−1
2 〈R◦g(z2)Gpq(z2)〉 − z−1
2
〈
R◦ 1
2n
G2
pq(z2)
〉
10 V. Vasilchuk
− z−1
2
2
(2n)2
〈
g◦(z1)g◦(z2)
((
G2(z1)G(z2)
)
pq
+
(
G(z2)G2(z1)
)
−q−p
)〉
− z−1
2
2
(2n)2
〈
(g◦(z1))
2 G3
pq(z2)
〉
.
Then we put p = q in all four cases and p = −q another time in the second case, and apply
1
2n
n∑
p=−n
and obtain
〈
|g◦(z)|4
〉
= 〈R◦g(z2)〉 = −z−1
2
〈
R◦g2(z2)
〉
− z−1
2
1
2n
〈
R◦ 1
2n
TrG2(z2)
〉
− z−1
2
4
(2n)2
〈
g◦(z1)g◦(z2)
1
2n
Tr
(
G2(z1)G(z2)
)〉
− z−1
2
2
(2n)2
〈
(g◦(z1))
2 1
2n
TrG3(z2)
〉
.
Using this relation, the bounds (26) and
∣∣〈R◦g2(z2)
〉∣∣ = |〈R◦g◦(z2)g(z2)〉+ 〈R◦g(z2)〉 〈g(z2)〉| ≤ 2
〈
|g◦(z)|4
〉
|Im z|
,∣∣∣∣〈R◦ 1
2n
TrG2(z2)
〉∣∣∣∣ ≤ v
|Im z|3
,
we obtain that for |Im z| ≥ 3
〈
|g◦(z)|4
〉
= O(n−3). Thus, due to the Schwartz inequality the
term
〈
g◦(z1) (g◦(z2))
2
〉
is of the order O(n−5/2) and, hence, r1,n is of the same order.
The cases of the terms rj,n, j = 2, . . . , 5 can be treated analogously, with exception for the
last term of r5,n. The last term of (20)
1
(2n)2
〈
1
2n
n∑
r,p=−n
(
G2(z1)
)
r−p
Grp(z2)
〉
can be treated as follows.
First, observe that in the case four 〈ĝ(z)〉 = o(n−1). Indeed, using (13) with q = −p, we
obtain
〈ĝ(z)〉 = −z−1 〈g(z)〉 〈ĝ(z)〉 − z−1 〈g◦(z)ĝ(z)〉 − z−1 1
2n
〈
1
2n
Tr
(
G(z)GT (z)
)〉
,
where GT is transpose of G. Due to the the Schwartz inequality for the trace, the last term in
r.h.s. of this relation is of the order O(n−1). Since the variance of g(z) is of the order O(n−2),
the second term is at least of the order O(n−1) (in fact it is of the order O(n−2), since, as one
can show, the variance of ĝ(z) is of the same order). Thus, 〈ĝ(z)〉 is of the order O(n−1). Its
easy to show in the same way that〈
ĥ(z)
〉
=
〈
1
2n
n∑
j=−n
(
G2(z)
)
j−j
〉
is also of the order O(n−1) and its variance is of the order O(n−2).
Now, using the resolvent identity for the average of
Φ =
1
2n
n∑
p,q=−n
(
G2(z1)
)
p−q
Gpq(z2),
On the GUEs with Additional Symmetries 11
relations (10) and (12), we obtain
〈Φ〉 = −z−1
2
〈
ĥ(z1)
〉
− z−1
2
1
(2n)3/2
〈
n∑
p,q,r,j,k=−n
(
G2(z1)
)
p−q
Grj(z2)
(
W ′
n
)
jk
Gkq(z2)
〉
− z−1
2
1
(2n)3/2
〈
n∑
p,q,r,j,k,m=−n
Gpj(z1)
(
W ′
n
)
jk
G(z1)kmGm−q(z1)Grq(z2)
〉
− z−1
2
1
(2n)3/2
〈
n∑
p,q,r,j,k,m=−n
Gpm(z1)G(z1)mj
(
W ′
n
)
jk
Gk−q(z1)Grq(z2)
〉
.
Substituting in this relation the value of W ′
n, we obtain
〈Φ〉 = −z−1
2
〈
ĥ(z1)
〉
− z−1
2 〈g(z2)Φ〉 − z−1
2
1
2n
〈
1
2n
n∑
p,q,r=−n
(
G2(z1)
)
p−q
Gr−p(z2)Grq(z2)
〉
− z−1
2
1
2n
〈
1
2n
n∑
p,q=−n
(G(z1)G(z2))pq G2
p−q(z1)
〉
− z−1
2
1
2n
〈
1
2n
n∑
p,q=−n
(
G2(z1)G(z2)
)
pq
Gp−q(z1)
〉
− z−1
2 〈ĝ(z1)Φ〉 − z−1
2
〈
ĥ(z1)Φ
〉
.
The first term of the r.h.s. is of the order O(n−1), the last five terms are of the same order,
because of the Schwartz inequality and of the bounds for the variances of ĝ(z1) and ĥ(z1). The
second term we rewrite as follows
−z−1
2 〈g(z2)Φ〉 = −z−1
2 〈g(z2)〉 〈Φ〉 − z−1
2 〈g◦(z2)Φ〉 ,
where due to the Schwartz inequality the last term is also of the order O(n−1). Thus, we conclude
that 〈Φ〉 is of the order O(n−1) and, hence, the last term of r5,n is of the order O(n−3). �
Proposition 4. In the third case we have〈
1
2n
TrG2(z)
〉
= −〈g(z)〉
z
1− z−2
1 + 〈g(z)〉 z−1 + z−2
+ o
(
n−2
)
.
Proof. Indeed, we have〈
1
2n
TrG2(z)
〉
=
〈
1
2n
Tr
d
dz
G(z)
〉
=
d
dz
〈g(z)〉 .
Hence, we can just take the derivative of the identity (14) for 〈g(z)〉. Since all terms of the order
O(n−2) in square brackets in (14) are analytical and uniformly bounded for |Im z1,2| ≥ 3, they
remain of the same order. Thus, we obtain the relation needed. �
Acknowledgements
The work of V.V. was supported by the Grant of the President of Ukraine for young scientists
GP/F8/0045. Author thankful to Professor L. Pastur for numerous helpful discussions.
12 V. Vasilchuk
[1] Bellissard J., Magnen J., Rivasseau V., Supersymmetric analysis of a simplified two-dimensional Anderson
model at small disorder, Markov Process. Related Fields, 2003, V.9, 1–30.
[2] Disertori M., Density of states for GUE through supersymmetric approach, Rev. Math. Phys., 2004, V.16,
N 9, 1191–1225.
[3] Disertori M., Rivasseau V., Random matrices and the Anderson model, math-ph/0310021.
[4] Khorunzhenko B., Khorunzhy A., Pastur L., Asymptotic properties of large random matrices with indepen-
dent entries, J. Math. Phys, 1996, V.37, 5033–5060.
[5] Khorunzhenko B., Khorunzhy A., Pastur L., Shcherbina M., Large-n limit in the statistical mechanics and
the spectral theory of disordered systems, in Phase Transitions and Critical Phenomena, Editors C. Dommb
and J. Lebowitz, London, Academic Press, 1992, 74–239.
[6] Khorunzhy A., Eigenvalue distribution of large random matrices with correlated entries, Mat. Fiz. Anal.
Geom., 1996, V.3, 80–101.
[7] Marchenko V.A., Pastur L.A., Distribution of eigenvalues for some sets of random matrices, Math. USSR
Sb., 1967, V.1, 457–483.
[8] Pastur L., Khorunzhy A., Vasilchuk V., On an asymptotic property of the spectrum of the sum of one-
dimensional independent random operators, Dopov. Nats. Akad. Nauk Ukrainy, 1995, N 2, 27–30 (in Rus-
sian).
[9] Schenker J.H., Schulz-Balde H., Semicircle law and freeness for random matrices with symmetries or corre-
lation, math-ph/0505003.
1 Introduction and main results
2 The limiting NCMs
3 The correlation functions
4 Conclusion
A Appendix
|