This preview shows page 1. Sign up to view the full content.
Unformatted text preview: D.S.G. POLLOCK: TOPICS IN TIMESERIES ANALYSIS
ALGEBRAIC POLYNOMIALS
Consider the equation φ0 + φ1 z + φ2 z 2 = 0. Once the equation has been divided
by φ2 , it can be factorised as (z − λ1 )(z − λ2 ) where λ1 , λ2 are the roots or
zeros of the equation which are given by the formula
(1) λ= −φ1 ± φ2 − 4φ2 φ0
1
.
2φ2 If φ2 ≥ 4φ2 φ0 , then the roots λ1 , λ2 are real. If φ2 = 4φ2 φ0 , then λ1 = λ2 .
1
1
If φ2 < 4φ2 φ0 , then the √
roots are the conjugate complex numbers λ = α + iβ ,
1
λ∗ = α − iβ , where i = −1.
There are three alternative ways of representing the conjugate complex
numbers λ and λ∗ :
λ = α + iβ = ρ(cos θ + i sin θ) = ρeiθ , (2) λ∗ = α − iβ = ρ(cos θ − i sin θ) = ρe−iθ , where
(3) ρ= α2 + β 2 and θ = tan−1 β
α . These are called, respectively, the Cartesian form, the trigonometrical form and
the exponential form.
The Cartesian and trigonometrical representations are understood by considering the Argand diagram: Im β λ θ
Re −θ α ρ λ*
Figure 1. The Argand Diagram showing a complex
number λ = α + iβ and its conjugate λ∗ = α − iβ . 1 ALGEBRAIC POLYNOMIALS
The exponential form is understood by considering the following series
expansions of cos θ and i sin θ about the point θ = 0:
cos θ = 1 −
(4) θ4
θ6
θ2
+
−
+ ··· ,
2!
4!
6! i sin θ = iθ − iθ5
iθ7
iθ3
+
−
+ ··· .
3!
5!
7! Adding these gives
(5) cos θ + i sin θ = 1 + iθ − θ2
iθ3
θ4
−
+
+ ···
2!
3!
4! = eiθ .
Likewise, by subtraction, we get
(6) θ2
iθ3
θ4
cos θ − i sin θ = 1 − iθ −
+
+
− ···
2!
3!
4!
= e−iθ . These are Euler’s equations. It follows from adding (5) and (6) that
(7) cos θ = eiθ + e−iθ
.
2 Subtracting (6) from (5) gives
−i iθ
(e − e−iθ )
2
1 iθ
= (e − e−iθ ).
2i sin θ =
(8) Now consider the general equation of the nth order:
φ0 + φ1 z + φ2 z 2 + · · · + φn z n = 0. (9) On dividing by φn , we can factorise this as
(z − λ1 )(z − λ2 ) · · · (z − λn ) = 0, (10) where some of the roots may be real and others may be complex. The complex
roots come in conjugate pairs, so that, if λ = α + iβ is a complex root, then
there is a corresponding root λ∗ = α − iβ such that the product (z − λ)(z − λ∗ ) =
z 2 − 2αz + (α2 + β 2 ) is real and quadratic. When we multiply the n factors
together, we obtain the expansion
(ii) λi z n−1 + 0 = zn −
i λi λj z n−2 − · · · (−1)n λ1 λ2 · · · λn .
i j 2 D.S.G. POLLOCK: TOPICS IN TIMESERIES ANALYSIS
This can be compared with the expression (φ0 /φn ) + (φ1 /φn )z + · · · + z n =
0. By equating coeﬃcients of the two expressions, we ﬁnd that (φ0 /φn ) =
(−1)n λi or, equivalently,
n (12) (−λi )−1 . φn = φ0
i=1 Thus we can express the polynomial in any of the following forms:
φi z i = φn
= φ0 (−λi )−1 = φ0 (13) (z − λi ) 1− z
λi (z − λi )
. We should also note that, if λ is a root of the primary equation φi z i = 0,
where rising powers of z are associated with rising indices on the coeﬃcients,
then μ = 1/λ is a root of the equation
φi z n−i = 0, which has declining
φi μ−i = 0 implies that
powers of z instead. This follows since
φi λi =
φi μn−i = 0. Confusion can arise from not knowing which of
μn φi μ−i =
the two equations one is dealing with.
Rational Functions of Polynomials
If β (z ) and φ(z ) are polynomial functions of z of degrees d and g respectively with d < g , then the ratio β (z )/φ(z ) is described as a proper rational
function. We shall often encounter expressions of the form
(14) y (t) = β (L)
x(t).
φ(L) For this to have a meaningful interpretation in the context of a timeseries
model, we normally require that y (t) should be a bounded sequence whenever
x(t) is bounded. The necessary and suﬃcient condition for the boundedness of
y (t), in that case, is that the series expansion of β (z )/φ(z ) should be convergent
whenever z  ≤ 1. We can determine whether or not the sequence will converge
by expressing the ratio β (z )/φ(z ) as a sum of partial fractions. The basic result
is as follows:
(15) If β (z )/φ(z ) = β (z )/{φ1 (z )φ2 (z )} is a proper rational function,
and if φ1 (z ) and φ2 (z ) have no common factor, then the function
can be uniquely expressed as
β (z )
β1 (z ) β2 (z )
=
+
,
φ(z )
φ1 (z ) φ2 (z )
where β1 (z )/φ1 (z ) and β2 (z )/φ2 (z ) are proper rational functions.
3 ALGEBRAIC POLYNOMIALS
Example. Consider the following partial fraction expansion
(16) gz + d
c1
c2
=
+
(1 − z/λ1 )(1 − z/λ2 )
(1 − z/λ1 ) (1 − z/λ2 ) To ﬁnd the value of c1 , the equation is multiplied on both sides by 1 − z/λ1 and
then z is set to λ1 . The result is that c1 = (gλ1 + d)/(1 − λ1 /λ2 ). Likewise, it
is found that c2 = (gλ2 + d)/(1 − λ2 /λ1 ).
Imagine that φ(z ) = (1 − z/λi ). Then, repeated applications of this
basic result enables us to write
(16) κ2
κg
κ1
β (z )
+
+ ··· +
.
=
φ(z )
1 − z/λ1
1 − z/λ2
1 − z/λg By adding the terms on the RHS, we ﬁnd an expression with a numerator of
degree n − 1. By equating the terms of the numerator with the terms of β (z ),
we can ﬁnd the values κ1 , κ2 , . . . , κg . The convergence of the expansion of
β (z )/φ(z ) is a straightforward matter. For the series converges if and only if
the expansion of each of the partial fractions converges. For the expansion
(17) κ
= κ 1 + z/λ + (z/λ)2 + · · ·
1 − z/λ to converge when z  ≤ 1, it is necessary and suﬃcient that λ > 1.
The expansion of (17) applies to complex roots as well as to real roots:
c
c∗
=c 1 + z/λ + (z/λ)2 + · · ·
+
1 − z/λ 1 − z/λ∗
+ c∗ 1 + z/λ∗ + (z/λ∗ )2 + · · · (18) ∞ = z t (cλ−t + c∗ λ∗−t ). t=0 The various complex quantities can be represented in terms of exponentials:
(19) λ = κ−1 e−iω , λ∗ = κ−1 eiω , = ρe−iθ , c∗ = ρeiθ . Then, the generic term of the expansion becomes
(20) z t (cλ−t + c∗ λ∗−t ) = z t ρκt ei(ωt−θ) + e−i(ωt−θ)
= z t 2ρκt cos(ωt − θ).
4 D.S.G. POLLOCK: TOPICS IN TIMESERIES ANALYSIS
The expansion converges for all z  ≤ 1 if and only if κ < 1, which is a
condition on the modulus of the complex number λ. But, κ = λ−1  = λ−1 ;
so it is conﬁrmed that the necessary and suﬃcient condition for convergence is
that λ > 1.
Finally, consider the case of a repeated root with a multiplicity of n. Then,
a Maclaurin series expansion is available that gives
(21) z
n(n + 1)
1
=1+n +
n
(1 − z/λ)
λ
2! z
λ 2 + n(n + 1)(n + 2)
3! z
λ 3 + ···. If λ is real, then λ > 1 is the condition for convergence. If λ is complex,
then it can be combined with the conjugate root in the manner of (20) to
create a trigonometric function; and, again, the condition for convergence is
that λ > 1.
This result can be understood by regarding the LHS of (21) as a representation of n transfer functions in series, each of which fulﬁls the BIBO condition.
The general conclusion is that the transfer function is stable if and only if
all of the roots of the denominator polynomial a(z ), which are described as the
poles of the transfer function, lie outside the unit circle in the complex plane.
The Expansion of a Rational Function: The Method of Detached
Coeﬃcients
The method of ﬁnding the coeﬃcients of the series expansion of the transfer
function in the general case can be illustrated by the secondorder case:
(22) β0 + β1 z
= ω0 + ω 1 z + ω2 z 2 + · · · .
1 − φ1 z − φ2 z 2 We rewrite this equation as
(23) β0 + β1 z = 1 − φ1 z − φ2 z 2 ω0 + ω1 z + ω2 z 2 + · · · . Then, by performing the multiplication on the RHS, and by equating the coeﬃcients of the same powers of z on the two sides of the equation, we ﬁnd
that (24) β0 = ω0 ,
β1 = ω1 − φ1 ω0 ,
0 = ω2 − φ1 ω1 − φ2 ω0 ,
.
.
.
0 = ωn − φ1 ωn−1 − φ2 ωn−2 , 5 ω0 = β0 ,
ω1 = β1 + φ1 ω0 ,
ω2 = φ1 ω1 + φ2 ω0 ,
.
.
.
ωn = φ1 ωn−1 + φ2 ωn−2 . ...
View
Full
Document
This note was uploaded on 03/02/2012 for the course EC 7087 taught by Professor D.s.g.pollock during the Fall '11 term at Queen Mary, University of London.
 Fall '11
 D.S.G.Pollock

Click to edit the document details