Chapter 3

Polynomials and partial fractions

3.1 Polynomials
3.1.1 De…nitions and notations
In this subsection, we shall give basic de…nitions, notations and vocabulary of poly-
nomials.
Let K be a …eld.

We denote by K(N) the set of sequences (an )n with elements in K such that

9p 2 N; 8n > p : an = 0:

An element of K(N) is called polynomial.

We de…ne on K (N) two binary operations: an addition and a multiplication as
follows:

(an )n + (bn )n = (an + bn )n :

and
X
n
(an )n (bn )n = (cn )n with cn = an b n k :
k=0

It is easy to verify that (K(N) ; +; :) is a commutative ring.

The neutral element for the addition is the polynomial (0; 0; 0; : : : ) called the
zero polynomial:

76
 The neutral element for the multiplication is the polynomial (1; 0; 0; : : : ):

An element of K(N) of the form

(a; 0; 0; : : : )

is called a constant polynomial.

The map
f K ! K(N)
a 7 ! (a; 0; 0; : : : )
is an epimorphism of rings which allows to identify (a0 ; 0; 0; : : : ) with a0 2 K
and so K can be considered as subring of K(N) : Then if a 2 K and (a0 ; a1 ; : : : ; am ; : : : ) 2
K(N) we have

a(a0 ; a1 ; : : : ; am ; : : : ) = (aa0 ; aa1 ; : : : ; aam ; : : : ):

We denote by X the polynomial

X = (0; 1; 0; 0; : : : ):

Direct calculations give

X 0 = (1; 0; 0; : : : ); X 2 = (0; 0; 1; 0; : : : ); X 3 = (0; 0; 0; 1; 0; : : : ):

Then we prove by induction on k 2 N that

X k = (0; : : : ; 0; 1 ; 0; 0; : : : ):
"
kth p o sitio n

Let P be an element of K(N) :

P = (a0 ; a1 ; : : : ; am ; 0; 0; : : : ) with am 6= 0:

Then P can be written as

P = (a0 ; 0; : : : ) + (a1 ; 0; : : : )X + (a2 ; 0; : : : )X 2 + + (am ; 0; : : : )X m :

Which gives right to the following canonical writing of a polynomial

X
m
2 m
P = a0 + a1 X + a2 X + + am X = ak X k :
k=0

77
 X X
Two polynomials P = ai X i and Q = bi X i are equal if and only if
i 0 i 0
an = bn for all n 2 N:
(N)
The ring (K
Pm; +; :) will be denoted K[X]. X is called the indeterminate and
when P = k=0 ak X k the ak ; 1 k m are called the coe¢ cients of P .
The elements of K[X] are denoted by capital letters

A; B; C; P; Q; R; : : :

Let P 2 K[X]:

– The degree of a nonzero polynomial P denoted by deg P is de…ned by:

deg P = maxfn 2 N j an 6= 0g. When deg P = p: The coe¢ cient ap is
called the leading coe¢ cient of P , the polynomial P is said to be monic
if ap = 1; and a0 is called the constant term of P:
– By convention degree of the zero polynomial is 1.
– The valuation of a nonzero polynomial P is de…ned by valP = minfn 2 N
j an 6= 0g.
– By convention valuation of the zero polynomial is +1.

Kn [X] denotes the set of polynomials on K of degree less than or equal to n.

It is easy to prove that Kn [X] is a subgroup of the group (K[X]; +):

Example 123

1. In R[X]; let P = 2 + 3X 2 4X 3 ; then deg P = 3; valP = 0; the leading

coe¢ cient of P is 4 and the constant term of P is 2:

2. Any nonzero constant polynomial is of zero degree. Conversely, any polynomial
of degree zero is constant.
3. In R[X]; let
P = (X + 1)n (X 1)n ; n > 1;
deg P = n 1; the leading coe¢ cient of P is 2n, the constant term of P is
1 ( 1)n :

78
 Pn
Let P = k=0 ak X k 2 K[X] (n 2 N): The map

Pe : KP! K
t 7 ! nk=0 ak tk ;

(with the convention 8t 2 K; t0 = 1) is called the polynomial map (function)

associated to P: It is clear that 8P; Q 2 K[X]; 8 2 K we have:

+ Q = Pe + Q:
1. P^ e

g
2. P Q = PeQ:
e

3. f
P = Pe:

I Let P; Q 2 K [X] such that for any t 2 K we have Pe(t) = Q(t): e This gives
(P^ Q)(t) = 0 8t 2 K. then, necessarily (P Q) = 0 (the only polynomial
possessing an in…nity roots is the zero polynomial) and so P = Q: This allows
us, when the …eld K is in…nite (this is particularly the case when K is a sub…eld
of C) to confound the polynomial P with the polynomial map Pe: In the next
example we will show that it is not the case when K is …nite.

Example 124

1. The polynomial map associated to the polynomial P = 1 + X 3 2 R [X] is the

map from R into R de…ned by Pe(t) = 1 + t3 :

2. Let P = X + X 3 2 F3 [X], then Pe(0) = Pe(1) = Pe(2) = 0, and so P is

identically zero as a map, but it is not the zero polynomial.

We give now some properties of the degree for that we agree to the convention:
1 + 1 = 1; a + 1 = 1; 1 < a for any a 2 Z.

Proposition 125 For any P , Q 2 K[X] and 2 K we have

1. deg(P + Q) 6 max(deg P; deg Q):

2. deg(P Q) = deg P + deg Q:

3. deg( P ) = deg P:

79
Proof

1. If P or Q is the zero polynomial, the inequality is obvious. Otherwise, let d

be the degree of A and e the degree of Q. Then P = a0 + + ad X d and
e
Q = b0 + + be X where ai and bi are in K. If d > e, then we can write:

P + Q = (a0 + b0 ) + + (ae + be )X e + (ae+1 )X e+1 + + ad X d

It is then clear that deg(P + Q) = d = max(deg P; deg Q). The case when
d < e is similar. Finally, when d = e, we have:

P + Q = (a0 + b0 ) + + (ad + bd )X d :

Either all coe¢ cients are zero, and deg(P + Q) = 1 making the inequality
obvious, or at least one is nonzero and the nonzero coe¢ cient with the highest
index is the leading coe¢ cient of P + Q, and its index is well less or equal to d:

2. We have
X
n+m
PQ = ck X k
k=0
Pk
where for any 0 k n + m; ck = i=0 ai bk i : Then, cn+m = an am 6= 0; so
cn+m (since an and am are not zero)) so is the leading coe¢ cient of P Q. Thus,

deg(P Q) = deg P + deg Q:

3. If 6= 0 and deg P = n; the leading coe¢ cient of P is an which is an 6= 0

since 6= 0 and an 6= 0. Then deg( P ) = n.

To give some properties of valuation we agree also to the convention: +1 + 1 =

+1; a + 1 = +1; a < 1 for any a 2 Z.

Proposition 126 For any P , Q 2 K[X] and 2 K we have:

1. val(P + Q) min(valP; valQ):

2. val(P Q) = valP + valQ:

3. val( P ) = valP:

80
Proof

1. It is easy to observe that if P or Q is the zero polynomial, we have the equality

val(P + Q) = min(valP; valQ): Suppose none of the two polynomials P and
Q is the zero polynomial and let m = valP and ` = valQ: We can write P =
am X m +am+1 X m+1 + +ad X d (am 6= 0) and Q = b` X ` +b`+1 X `+1 + +be X e
(b` 6= 0): If ` < m then, P + Q = a` X ` + a`+1 X `+1 + ; it is clear then that
val(P + Q) = ` = min(valP; valQ). The case m < ` is similar. Finally, when
m = ` we have P + Q = (am + bm )X m + (am+1 + bm+1 )X m+1 + Either
the coe¢ cients are zero and val(P + Q) = +1 making the inequality obvious.
Otherwise at least one coe¢ cient is nonzero. If the nonzero coe¢ cient with
least index is at + bt then val(P + Q) = t which is well greater or equal to m.

2. If P or Q is the zero polynomial, it is obvious that we have equality. Suppose

that none of the two polynomials is zero, let m be the valuation of P and ` the
valuation of Q, then P = ad X d + + am and Q = be X e + + b` (am 6= 0 and
b` 6= 0 ). Then we can write: P Q = am b` X m+` + + ad be X d+e with am b` 6= 0;
then val(P Q) = m + ` = valP + valQ:

3. Let 6= 0 and P = a0 + + ad X d with valP = m: That means that m is the

least n 2 N such that an 6= 0: That is P = am X m + am+1 X m+1 + + ad X d
and am 6= 0: Since 6= 0 then P = am X m + am+1 X m+1 + + ad X d with
am 6= 0. Then val( P ) = m.

Remark 127 Let P be a nonzero polynomial in K[X]: The following are obvious

1. valP deg P .

2. val0 deg 0:

3. P is a constant polynomial if and only if valP = deg P:

Theorem 128 The ring K[X] is an integral domain and its only invertible elements
are nonzero constant polynomials.

Proof Let P 2 K[X] be an invertible polynomial, then there exists Q 2 K[X] such
that P Q = 1. Note that neither P nor Q can be zero. Taking the degree, we obtain

deg(P Q) = deg P + deg Q = deg 1 = 0:

81
Since deg P and deg Q are non negative integers and their sum is zero, necessarily

deg P = deg Q = 0:

This is equivalent to say P and Q are nonzero constant polynomials.

3.1.2 Composition of polynomials

X
n
Let P and Q be two polynomials of K [X] with P = ak X k : We denote by P Q
i=0
X
n
the polynomial de…ned as P Q= ai Qk ; where it is always agreed that Q0 = 1:
k=0
The polynomial P Q is referred to as the polynomial composed of P and Q. It
is denoted P Q or P (Q) and is said to have been obtained by substitution of the
polynomial Q with indeterminate X in P (X):

Example 129 In R[X]; let P = X 3 + X 2 + X + 1 and Q = 2X 2 + 1: We have

P (Q) = (2X 2 + 1)3 + (2X 2 + 1)2 + (2X 2 + 1) + 1 = 8X 6 + 20X 4 + 8X 2 + 4:

Q(P ) = 2(X 3 + X 2 + X + 1)2 + 1 = 2X 6 + 4X 5 + 6X 4 + 8X 3 + 6X 2 + 4X + 3:

Proposition 130 Let P; Q; R 2 K[X] and 2 K.

1. The composition of polynomials is not commutative.

2. Composition of polynomials is associative ((P Q) (R) = P (Q R)):

3. X P =P X = P:

4. (P Q) R = (P Q)(Q R):

5. (P + Q) R = (P R) + ( Q R):

6. When P; Q are di¤erent from the zero polynomial we have deg (P Q) =

deg P deg Q:

82
3.1.3 Euclidean algorithm (Long division)
Theorem 131 For any A; B 2 K[X] with B =
6 0, there exists a unique pair (Q; R) 2
K[X]2 such that
A = BQ + R with deg R < deg B: (3.1)

The polynomials Q and R are called, respectively, the quotient and remainder of
the Euclidean division of A by B. The polynomial B is said to divide A or A is said
to be divisible by B if R = 0:
Proof
Existence: We prove the result by induction on the degree of A. When deg A <
deg B, we take (Q; R) = (0; A). Let us assume that deg A deg B and suppose that
the property is true for any polynomial P such that deg A n and show it for a
polynomial A with deg A = n + 1. Put
an+1 n+1 m
A1 = A X B;
bm
where an+1 and bm are the leading coe¢ cients of A and B, respectively, then
deg(A1 ) n. Applying the induction hypothesis to A1 implies that there exists
a pair (Q1 ; R1 ) of polynomials in K[X] such that

A1 = Q1 B + R1 ; deg R1 < deg B;

then
an+1 n+1 m
A= Q1 + X B + R1 :
bm
Uniqueness: Suppose (Q1 ; R1 ) and (Q2 ; R2 ) are two pairs of polynomials satisfying
(3.1), then
R2 R1 = B(Q1 Q2 ):
Then
deg(R2 R1 ) = deg(B(Q1 Q2 )) = deg B + deg(Q1 Q2 ):
As deg(R2 R1 ) < deg B necessarily Q1 Q2 = 0 and R2 R1 = 0; which gives
Q1 = Q2 and R2 = R1 :

Example 132

1. Every polynomial divides the zero polynomial.

2. X 2 + X + 1 divides X 6 1:

83
 3. For any n 2 N; X 1 divides X n 1.

Proposition 133

1. If 2 K ; then divides any A 2 K[X]:

2. Let A; B 2 K[X] and B 6= 0: If B divides A; then deg A deg B:

Proof

1. Since 2 K we have P = ( 1 P ) then, the result follows.

2. If B divides A; we have A = BQ then, deg A = deg B + deg Q. Hence deg A

deg B:

Example 134 In R[X]; let

A = 4X 5 3X 4 + 5X 2 + 1

and
B = X 3 + 1:
We have

4X 5 3X 4 + 5X 2 + 1 X3 + 1
2
4X 5 + 4X 2 4X
| {z 3X}
Q

3X 4 + X 2 + 1
3X 4 3X
2
+X
| +{z3X + 1}
R

Then
5 4
4X 3X{z+ 5X 2 + 1} = (X 3 + 1) (4X 2 3X) + X 2
+ {z
3X + 1} :
| | {z } | {z } |
A B Q R

84
3.1.4 Division by increasing power order
Theorem 135 Let A; B 2 K[X] with nonzero constant term of B and let n be a
positive integer. Then there exists a unique pair (Q; R) in K[X]2 such that

A = BQ + X n+1 R with deg Q n:

The polynomials Q and R are called, respectively, the quotient and remainder of
the division according to increasing powers of A by B, and n is called the order of
the division.
Proof
Existence: By induction on n. For n = 0, let a0 and b0 be the constant terms
of A and B respectively. Then the constant term of the polynomial
a0
A B
b0
is zero, so
a0
A B = XR; with R 2 K[X]:
b0
a0
Putting Q = b0
we have

A = BQ + XR with deg Q n:

Assume that the property is satis…ed for n, then there exists a pair of polynomials
(Q; R) 2 K[X]2 such that

A = BQ + X n+1 R with deg Q n:

Since the property is veri…ed for n = 0, we can write R as follows:

R = BQ1 + XR1 ;

then

A = B(Q + X n+1 Q1 ) + X n+2 R1 with deg(Q + X n+1 Q1 ) n + 1:

Uniqueness: Suppose

A = BQ + X n+1 R (1) with deg Q n (3.2)

and
A = BQ1 + X n+1 R1 (2) with deg Q1 n (3.3)

85
 Calculating (3:2)-(3:3) we get

B(Q Q1 ) = X n+1 (R1 R):

Applying the valuation, we get
val(B(Q Q1 )) = val(X n+1 (R1 R)):
Then
valB + val(Q Q1 ) = valX n+1 + val(R1 R)
valB
| {z } + val(Q Q ) = n + 1 + val(R1 R) : (3.4)
| {z 1} | {z }
=0

Since deg(Q Q1 ) n and the valuation of a polynomial is less or equal than

its degree (except if P is the zero polynomial). Then (3.4) is not possible only if
val(R1 R) = val(Q Q1 ) = +1;
which is equivalent to Q Q1 = 0 and R1 R = 0; consequently Q = Q1 and
R1 = R; which is required.
Example 136 Let us calculate in R [X] the increasing power division to order 3 of
A by B where
A = 1 + X and B = 1 + X 2 :

1+X 1 + X2
1 + X2 1| + X {zX 2 X}3
Q

X X2
X + X3
X2 X3
X2 X4
X3 + X4
X3 X5
X 4 + X 5 = X 4 (1 + X)
| {z }
R

Which gives (1 + X) = (1 + X ) (1 + X X 2
2
X 3 ) +X 4 (1 + X) :
| {z } | {z } | {z } | {z }
A B Q R

86