222 lines
5.7 KiB
Typst
222 lines
5.7 KiB
Typst
#import("@local/dtu-template:0.5.1"):*
|
|
|
|
#show: dtu-note.with(
|
|
course: "01017",
|
|
course-name: "Discrete Mathematics",
|
|
title: "Polynomials and the Extended Euclidean Algorithm",
|
|
date: datetime(year: 2025, month: 12, day: 4),
|
|
author: "Rasmus Rosendahl-Kaa",
|
|
semester: "Fall 2025"
|
|
)
|
|
An example of a polynomial
|
|
$
|
|
f(x) = x^2 - 4x+3\
|
|
g(x)= 2x-3\
|
|
h(x) = 7
|
|
$
|
|
The curve on the graph is called a parabola
|
|
|
|
What we say for polynomials with real coefficients also applies to the ones with complex coefficients
|
|
|
|
#definition(title: "Polynomial of degree n")[
|
|
$
|
|
P(x)=a_0 + a_1 x + a_2 x^2 + dots + a_n x^n, quad a_n eq.not 0, a_i in RR "or" CC
|
|
$
|
|
- $a_n$ is called the leading term
|
|
- $a_0$ is called the constant term.
|
|
- $n$ is the degree
|
|
]
|
|
|
|
#definition(title: "Addition of polynomials")[
|
|
Same $P(x)$ as before
|
|
$
|
|
Q(x) = b_0 + b_1 x + dots + b_m x^m, quad m <= n\
|
|
$
|
|
|
|
$
|
|
P(x) + Q(x) = (a_0 + b_0) + (a_1 + b_1) x + dots + (a_m + b_m)x^m\ + a_(m+1)x^(m+1) + dots + a_n x^n
|
|
$
|
|
$
|
|
deg(P(x) + Q(x)) <= n "with equality if" m < n
|
|
$
|
|
]
|
|
|
|
#definition(title: "Multiplication")[
|
|
Same $P(x), Q(x)$ as before
|
|
$
|
|
P(x) dot Q(x) = a_0 dot b_0 + (a_0 b_1 + a_1 b_0) x + (a_0 b_2 + a_1 b_1 + a_2 b_0) x^2 + dots +\
|
|
a_n b_m x^(n+m)
|
|
$
|
|
$
|
|
deg(P(x) dot Q(x)) = n+m = deg(P(x)) + deg(Q(x))
|
|
$
|
|
When multiplying you basically do:
|
|
$
|
|
P(x) dot Q(x) = a_0 dot Q(x) + a_1x dot Q(x) + dots + a_n x^n Q(x)
|
|
$
|
|
]
|
|
|
|
== Divisible
|
|
#definition(title: "Divisible")[
|
|
$M(x)$ divides $N(x)$ (we write $M(x) | N(x)$) if $N(x) = Q(x) dot M(x)$
|
|
|
|
$Q(x)$ is some polynomial.
|
|
|
|
We have: $deg(N) = deg(Q) + deg(M)$, so $deg(M) <= deg(N)$
|
|
|
|
So basically, you need to find a polynomial $Q(x)$ so that $Q(x) dot M(x) = N(x)$, then $M(x)$ divides $N(x)$
|
|
|
|
If $M(x) | N(x)$ and $N(x)|M(x)$, then they must have the same degree. And then $deg(Q)$ must have degree 0 and be a constant.
|
|
|
|
$exists alpha in RR: N(x) = alpha dot M(x)$
|
|
]
|
|
|
|
=== Common divisor
|
|
$D(x)$ is a common divisor of $M(x), N(x)$ if $D(x) | M(x) "and" D(x) | N(x)$
|
|
|
|
=== Greatest common divisor
|
|
#definition()[
|
|
$D_1(x)$ is _a_ greatest common divisor (gcd) of $M(x), N(x)$ if and only if $D$ is a common divisor and $D_1(x)$ also satisfies:
|
|
$ (D_1(x) | M(x) and D_1(x) | N(x)) => D_1(x) | D(x) $
|
|
|
|
If $D_1(x)$ is a greatest common divisor, then $D_1(x)$ times a constant is also a greatest common divisor
|
|
|
|
Suppose $D_2(x)$ is also a $"gcd"(M(x), N(x))$, then $D_2(x) | D_1(x)$ and $D_1(x) | D_2(x)$ so $D_2 = alpha D_1$
|
|
]
|
|
|
|
#note-box()[
|
|
There can be more than one greatest common divisor
|
|
]
|
|
|
|
Given $N(x), M(x)$, find a gcd.
|
|
|
|
#note-box(title: "For integers (repetition)")[
|
|
|
|
For integers: $n, m,$ find $"gcd"(n,m)$.
|
|
==== Euclid
|
|
$ n &= q_1 dot m + r_1, quad 0 <= r_1 < r_0 = m\
|
|
r_0 &= q_2 dot r_1 + r_2\
|
|
r_1 &= q_3 dot r_2 + r_3\
|
|
dots.v\
|
|
r_(k-3)&= q_(k-1) dot r_(k-2) + r_(k-1)\
|
|
r_(k-2)&= q_k dot r_(k-1) + r_k\
|
|
r_(k-1)&= q_(k+1) dot r_(k) + 0
|
|
$
|
|
$r_k$ is the greatest common divisor.
|
|
|
|
*Why is it a divisor*
|
|
|
|
It is the divisor because looking at the last line:
|
|
$r_k$ divides $r_(k-1), r_k$
|
|
|
|
We can go a line up: $r_(k-1)$ divides $r_(k-2), r_(k-1)$, but $r_k$ must also divide them.
|
|
|
|
Can go up a line again: $r_k$ divides $r_(k-3), r_(k-2)$ up until we get $r_k$ divides $n, m$
|
|
|
|
*Why is it the greatest common divisor*
|
|
|
|
$r_k$ can be written as a linear combination of $r_(k-2) "and" r_(k-1)$ which coefficients are integers.
|
|
|
|
You can go a line up and write $r_(k-1)$ as a linear combination, which you can input into $r_k$'s linear combination. Continue until you get:\
|
|
$
|
|
r_k = A dot N + B dot M
|
|
$
|
|
]
|
|
#definition(title: "GCD for polynomials")[
|
|
$deg(M) = m < n = deg(N)$
|
|
$
|
|
N(x) &= Q_1(x) dot M(x) + R_1(x), quad deg(R_1) < deg(M)\
|
|
M(x) &= Q_2(x) dot R_1(x) + R_2(x), quad deg(R_2) < deg(R_1)\
|
|
&dots.v\
|
|
R_(k-2)(x)&=Q_k (x) dot R_(k-1)(x)+R_k (x), quad deg(R_k) = 0\
|
|
R_(k-1)(x)&=Q_(k+1)(x) dot R_(k) (x)+0
|
|
$
|
|
|
|
$R_k (x) = A(x) dot N(x) + B(x) dot M(x)$
|
|
|
|
$A(x), B(x)$ are some polynomials.
|
|
|
|
$deg(R_k (x)) = deg(N(x)) - deg(M(x))$
|
|
|
|
]
|
|
|
|
#example()[
|
|
Find the greatest common divisor of
|
|
$
|
|
N(x) &= x^4 + x^3 - 2x^2 + 2x - 2 "and"\
|
|
M(x) &= x^2 + 2x -3
|
|
$
|
|
|
|
Divide:
|
|
|
|
$
|
|
underline(x^2+2x-3 |) x^4+3x^3-2x^2+2x-2 &underline(| x^2-x+3)\
|
|
underline(x^4 + 2x^3 -3x^2) &\
|
|
-x^3+x^2+2x-2&\
|
|
underline(-x^3-2x^2+3x)&\
|
|
3x^2-x-2&\
|
|
underline(3x^2+6x-9)&\
|
|
-7x+7
|
|
$
|
|
|
|
So:
|
|
$
|
|
N(x)=(x^2-x+3)M(x) + (-7x+7)
|
|
$
|
|
|
|
Now continue with the two new polynomials you found:
|
|
$
|
|
underline(-7x+7|) x^2+2x-3 &underline(|-1/7x-3/7)\
|
|
underline(x^2 - x)\
|
|
3x-3\
|
|
underline(3x-3)\
|
|
0
|
|
$
|
|
|
|
So:
|
|
$
|
|
M(x) = (-1/7 x - 3/7) dot (-7x+7) + 0
|
|
$
|
|
|
|
Now we're finished as we have 0. The greatest common divisor is $-7x+7$
|
|
We can write:
|
|
$
|
|
D(x)=-7x +7 = N(x) - (x^2-x+3) dot M(x)
|
|
$
|
|
|
|
To find $D_1(x)$ (a divisor of $D(x)$:
|
|
Remember: $D_1(x) = D(x) dot alpha$ where $alpha$ is a constant
|
|
$
|
|
D_1(x) = -x+1
|
|
$
|
|
#note-box()[
|
|
Both $D(x)$ and $D_1(x)$ are greatest common divisors of $N(x), M(x)$. as $D(x) = 1 dot D(x)$ (constant here is just $1$).
|
|
]
|
|
]
|
|
|
|
= Roots of polynomials
|
|
For the polynomial $a x^2 + b x + c$, the roots are: $(-b plus.minus sqrt(b^2 - 4 a c))/(2 a)$
|
|
|
|
Let's assume $"gcd"(N(x), M(x)) = D(x)$ then $alpha$ is a common root of $N(x), M(x) <=> alpha$ is a root in $D(x)$
|
|
|
|
$
|
|
N(x)=D(x) dot Q_1 (x)\
|
|
M(x)=D(x) dot Q_2 (x)
|
|
$
|
|
If $alpha$ is a root in $D(x)$, then it must also be a root in $M(x)$ and $N(x)$. The reason is that we can write $N, M$ as above
|
|
|
|
$
|
|
D(x) = A(x) N(x) + B(x) M(x)
|
|
$
|
|
|
|
$alpha$ is a root of $P(x) <=> (x- alpha)| P(x)$ which means $exists Q(x): P(x) = Q(x)(x-alpha) + beta$ where $beta$ is a constant
|
|
|
|
We can find $beta$ by calculating $P(alpha)$:
|
|
$
|
|
P(alpha) = Q(alpha)(alpha-alpha) + beta\
|
|
P(alpha) = Q(alpha)(0) + beta\
|
|
P(alpha) = beta
|
|
$
|
|
|
|
$x^2 +1= (x-i) dot (x+i)$
|