Covariance of The Number of Real Zeros of A Random Trigonometric Polynomial

COVARIANCE OF THE NUMBER OF REAL ZEROS OF
A RANDOM TRIGONOMETRIC POLYNOMIAL

K. FARAHMAND AND M. SAMBANDHAM
Received 8 March 2006; Accepted 26 March 2006
For random coecients

a j and b j we consider a random trigonometric polynomial de
fined as Tn () = nj=0 {a j cos j + b j sin j }. The expected number of real zeros of Tn ()
in the interval
(0,2) can be easily obtained. In this note we show that this number is in
fact n/ 3. However the variance of the above number is not known. This note presents a
method which leads to the asymptotic value for the covariance of the number of real
zeros of the above polynomial in intervals (0, ) and (,2). It can be seen that our
method in fact remains valid to obtain the result for any two disjoint intervals. The applicability
of our method to the classical random trigonometric polynomial, defined as

Pn () = nj=0 a j ()cos j, is also discussed. Tn () has the advantage on Pn () of being
stationary, with respect to , for which, therefore, a more advanced method developed
could be used to yield the results.
Copyright 2006 Hindawi Publishing Corporation. All rights reserved.
1. Introduction
Let (,Pr,) be a fixed probability space and for let {a j ()}nj=0 and {b j ()}nj=0 be
sequences of independent, identically and normally distributed random variables, both
with means zero and variances one. Denote by Nn (,) the number of real zeros of random trigonometric polynomial
Tn (,) Tn () =
n

a j ()cos j + b j ()sin j
(1.1)
j =0
in the interval (,) and by ENn (,) its expected value. Indeed the above definition of
random trigonometric polynomials diers from the classical case of
Pn (,) Pn () =
n
a j ()cos j
j =0
Hindawi Publishing Corporation

International Journal of Mathematics and Mathematical Sciences
Volume 2006, Article ID 28492, Pages 16
DOI 10.1155/IJMMS/2006/28492
(1.2)
Covariance of the number of real zeros
which has been extensively studied. The literature includes the original work of Dunnage
[3] which was later extended by Das [2] and Wilkins [8] and was reviewed by BharuchaReid and Sambandham [1] and recently by Farahmand [7]. They generally show that for
all suciently large n and for dierent classes of distributions of the coecients or in
dierent cases, forexample, the level crossing case instead of zero crossings, ENn (0,2) is
asymptotic to 2n/ 3. In particular the above work of Wilkins is of interest as it shows that
the error term involved in the asymptotic estimate is small and in fact is O(1). However,
finding the variance of the number of real zeros involves a dierent level of diculties.
There have been several attempts, for instance, see [4] or [6], to obtain the asymptotic
value for the variance of Nn (0,2) for Pn (). So far, the results are only in the form of
upper bounds. As far as the expected number of zeros is concerned the asymptotic value
of ENn (0,2) for Tn () and Pn () is the same. Therefore, we conjecture that their variances are also the same. In addition, with the above assumptions of independence of the
coecients a j () and b j () the inner term of Tn () given in (1.1) has the property of
being stationary with respect to . This can be seen by evaluating its covariance function
as

E a j ()cos j + b j ()sin j a j ()cos j( + ) + b j sin j( + )

= cos j cos j( + ) + sin j sin j( + ) = cos j.
(1.3)
Therefore it is natural to seek to evaluate the variance of number of zeros of Tn () which

possess the above stationary property instead of Pn () given in (1.2). We, however, are
unable to make any substantial progress in this direction. Instead, we obtain the covariance of the number of real zeros in the intervals (0, ) and (,2). As our main aim
remains to estimate the variance of Nn (0,2) we will present our results and discussions
in such a way that they could be used to be generalized for variance. Although we are
considering two intervals (0, ) and (,2) our proof is also valid for any two disjoint
intervals. A small modification and some generalization to our analysis should lead to an
asymptotic value for the variance. Looking at our proof it suggests that our estimate for
the covariance will remain the same as for the variance. Furthermore, although we are
considering the polynomial Tn () given in (1.1) as far as the results for the covariance,
and, therefore, the variance are concerned, it should remain invariant also for Pn (). For
random trigonometric polynomial Tn () given in (1.1) we prove the following.
Theorem 1.1. With the above assumption of independent and Gaussian distribution of the
coecients {a j ()}nj=0 and {b j ()}nj=0 the covariance of the number of real zeros of Tn () is
cov Nn (0,),Nn (,2) = 4n + O(1).
(1.4)
2. Covariance of the number of real zeros

For any two intervals (,) and (,) it is known that
E Nn (,)Nn (,) =

|xy | p1 ,2 (0,0,x, y)dx dx d1 d2 ,
(2.1)
K. Farahmand and M. Sambandham 3

where p1 ,2 (z1 ,z2 ,x, y) denotes the four-dimensional joint probability density function
of Tn (1 ), Tn (2 ), Tn (1 ), and Tn (2 ). For our purpose and using the above formula
to obtain the result for The covariance case, the two intervals (,) and (,) are disjoint. However the above formula and the following discussions remain valid for any two
intervals, whether or not they are overlapping. Let be the 4 4 variance-covariance
matrix of random variables Tn (1 ), Tn (2 ), Tn (1 ), and Tn (2 ) with cofactor i j of i jth
element. Then using the Gaussian assumption for the coecients we can calculate the
above-required joint density function as
33 x2 + 44 y 2 + 34 + 43 xy
exp
.
P1 ,2 (0,0,x, y) = 2
2||
4 ||
1
(2.2)
In order to evaluate (2.1) further in (2.2) we let q = x 33 / || and s = y 44 / ||. As

we will see later 33 and 44 are positive and therefore q and s are real. Hence from (2.2)
we obtain

=
=
|xy | p1 ,2 (0,0,x, y)dx dx

||3/2
4 2 33 44

||3/2
q2 + s2
34 + 43
|qs| exp

qs dq ds
2
33 44
4 2 33 44
(2.3)

q2 + s2 + 2qs
|qs| exp
dq ds,
where = (34 + 43 )/2 33 44 . Now let = cos, then from [7, page 97] the integral
appear in (2.3) can be evaluated as 4{1 + (/2 )cot}/ cos2 . Therefore for Gaussian
assumption the required formula in (2.1) is simplified as
1
ENn (,)Nn (,) = 2
Now we let

||3/2 1 + (/2 )cot
33 44 cos2
An 1 ,2 = cov Tn 1 ,Tn 2 ,

d1 d2 .
Cn 1 ,2 = cov Tn 1 ,Tn 2 ,

(2.4)
Bn 1 ,2 = cov Tn 1 ,Tn 2 ,
(2.5)
where Tn () is the derivative of Tn () with respect to . It is easy to show that the
cov{Tn (),Tn ()} = 0, also An (,) = var{Tn ()} = n and Bn (,) = var{Tn ()} = n(n +
1)(2n + 1)/6 are independent of . Therefore we can obtain the variance-covariance matrix of random variables Tn (1 ), Tn (2 ), Tn (1 ), and Tn (2 ) as

An 1 ,2
Cn 1 ,2
An 1 ,2
n

Cn 1 ,2

Cn 1 ,2
n(n + 1)(2n + 1)
6

Bn 1 ,2
Cn 1 ,2

.
Bn 1 ,2
n(n + 1)(2n + 1)
(2.6)
Let
sin (2n + 1) 1 2 /2

,
sin 1 2 /2
cos (2n + 1) 1 2 /2

.
=
sin{ 1 2 /2}
Sn 1 ,2 =
Zn 1 ,2
(2.7)
Then we can obtain the remaining elements of the above matrix as

An 1 ,2 An 2 ,1 =
n
j =0
Cn 1 ,2 Cn 2 ,1 =
n
Sn 1 ,2 1
,
2
cos j 1 ,2 =

j sin j 1 2
j =0
Sn (1 ,2 )cot 1 2 /2
(2n + 1)Zn 1 ,2
+
,
=
4
4

Bn 1 ,2 Bn 2 ,1 =
n
j 2 cos j 1 2
j =0
Sn 1 ,2

(2n + 1)2
(2n + 1)
Sn 1 ,2
+
Zn 1 ,2 cot
8
8
4

Sn 1 ,2
cot2
4
1 2
2
1 2
2
.
(2.8)
Now we are in the position to proceed with the proof of our theorem.
3. Proof of the theorem
From (2.6) we can obtain the determinate of as
|| = n2
n(n + 1)(2n + 1)
6
2

n2 Bn2 1 ,2

n(n + 1)(2n + 1)

n(n + 1)(2n + 1) 2
2nCn2 1 ,2
A2n 1 ,2

(3.1)

+ A2n 1 ,2 Bn2 1 ,2 + 2An 1 ,2 C 2 1 ,2 Bn 1 ,2 + Cn4 1 ,2 .

Also the required cofactors are

43 = 34 = n2 Bn 1 ,2 A2 1 ,2 Bn 1 ,2 A 1 ,2 C 2 1 ,2 ,
33 = 44 =

n(n + 1)(2n + 1) 2

n3 (n + 1)(2n + 1)
n2 C 2 1 ,2 +
An 1 ,2 .
6
6
(3.2)
(3.3)
K. Farahmand and M. Sambandham 5

Now we use the advantage that 1 and 2 are disjoint and therefore Sn (1 ,2 ), Zn (1 ,2 )
and cot(1 2 ) are bounded. Therefore from (3.1)(3.3) we obtain
|| n2
|33 |
n(n + 1)(2n + 1)
6
2
n3 (n + 1)(2n + 1)
,
6
(3.4)

|34 | 43 n2 Bn 1 ,2 = O n4 Sn 1 ,2 .
In order to evaluate the integral that appears in (2.4) we note that from (3.4)
+ 43
= 34
0
2 33 44
(3.5)
as n and therefore for suciently large n,

= arccos
.
2
(3.6)
This summarizes the value of (2.4) to

E Nn (,)Nn (,) =

||3/2
2 33 44
d1 d2 .
(3.7)
With our assumptions of our theorem it turns out that ||, 33 , and 44 are independent
of 1 and 2 . Also for all suciently large n,
||
n4 (n + 1)2 (2n + 1)2 n8 n7
+ ,
36
9
3
(3.8)
n3 (n + 1)(2n + 1) n5
.
44 33
6
3
Therefore
E{Nn (0,)Nn (,2)}
||3/2
33 44
n8 /9 + n7 /3
(n5 /3)2
3/2
n2 9n
+ .
3
2
(3.9)
In order to proceed we need to find ENn (0,) and ENn (,2). To this end we use the
Kac-Rice formula and because of the stationary property of Tn () mentioned above, we
are able to obtain an estimate with small error easily. Using a same method as [5] and
since cov(Tn (),T ()) = 0, we have
1
ENn (0,) =
where

0
B
d,
A
(3.10)
A2 = var Tn () = n,
B 2 = var Tn () =
n(n + 1)(2n + 1)
.
6
(3.11)
Therefore

n(n + 1)(2n + 1)
ENn (0,) =
=
6n

n2 n 1
n
1
3
.
+ + = +
+O
3 2 6
4
n
3
(3.12)
Hence by (3.9), (3.12) and since a similar result to (3.12) can be obtained for ENn (,2),
we can obtain
cov Nn (0,),Nn (,2)
n2 9n
n
3
+
+
+ O(1)
3
2
4
3
2
4n + O(1).
(3.13)
This completes the proof of the theorem.

References
[1] A. T. Bharucha-Reid and M. Sambandham, Random Polynomials, Probability and Mathematical
Statistics, Academic Press, Florida, 1986.
[2] M. Das, The average number of real zeros of a random trigonometric polynomial., Proceedings of
the Cambridge Philosophical Society 64 (1968), 721729.
[3] J. E. A. Dunnage, The number of real zeros of a random trigonometric polynomial, Proceedings of
the London Mathematical Society 16 (1966), 5384.
[4] K. Farahmand, On the variance of the number of real roots of a random trigonometric polynomial,
Journal of Applied Mathematics and Stochastic Analysis 3 (1990), no. 4, 253261.
, Number of real roots of a random trigonometric polynomial, Journal of Applied Mathe[5]
matics and Stochastic Analysis 5 (1992), no. 4, 307313.
, On the variance of the number of real zeros of a random trigonometric polynomial, Journal
[6]
of Applied Mathematics and Stochastic Analysis 10 (1997), no. 1, 5766.
, Topics in Random Polynomials, Pitman Research Notes in Mathematics Series, vol. 393,
[7]
Longman, Harlow, 1998.
[8] J. E. Wilkins Jr., Mean number of real zeros of a random trigonometric polynomial, Proceedings of
the American Mathematical Society 111 (1991), no. 3, 851863.
K. Farahmand: Department of Mathematics, University of Ulster at Jordanstown, Co. Antrim,
BT37 0QB, UK
E-mail address: k.farahmand@ulster.ac.uk
M. Sambandham: Department of Mathematics, Morehouse College, Atlanta, GA 30314, USA
E-mail address: msamband@morehouse.edu
Advances in
Operations Research
http://www.hindawi.com
Volume 2014
Advances in
Decision Sciences
Volume 2014
Journal of
Applied Mathematics
Algebra


Volume 2014
Journal of
Probability and Statistics

Volume 2014
The Scientific
World Journal

Volume 2014
International Journal of
Differential Equations
Volume 2014
Volume 2014
Submit your manuscripts at

Advances in
Combinatorics
Mathematical Physics
Volume 2014
Journal of
Complex Analysis
Volume 2014
International
Journal of
Mathematics and
Mathematical
Sciences
Mathematical Problems
in Engineering
Journal of
Mathematics
Volume 2014

Volume 2014
Volume 2014

Volume 2014
Discrete Mathematics
Journal of
Volume 2014

Discrete Dynamics in
Nature and Society
Journal of
Function Spaces
Abstract and
Applied Analysis
Volume 2014

Volume 2014

Volume 2014
Journal of
Stochastic Analysis
Optimization


Volume 2014
Volume 2014

Covariance of The Number of Real Zeros of A Random Trigonometric Polynomial

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Covariance of The Number of Real Zeros of A Random Trigonometric Polynomial

Hochgeladen von

Copyright:

Verfügbare Formate

COVARIANCE OF THE NUMBER OF REAL ZEROS OF

A RANDOM TRIGONOMETRIC POLYNOMIAL

For random coecients

Hindawi Publishing Corporation