Vous êtes sur la page 1sur 4

Now with, with complex numbers we

achieved to translation by addition.


So in order to do the same thing with
vectors, we have to introduce a vector
addition operation.
So here is the basic definition of vector
addition.
To add two vectors, you just add them
entry wise.
Here is the, here is the code for, doing
addition of two vectors, in Python, when
the two vectors are represented as lists.
Now there's a, always a 0 vector, and
notice that adding the 0 vector to any
vector gives you back that vector.
Addition has these various algebraic
properties, associativity and
communitivity.
And we'll see how those get used.
Now like complex numbers, we can
represent vectors as arrows in the plane.
So this Vector of 3, 1, 5 can be
represented with its tail at the origin
and its head here.
It could equally well be represented with
its tail over here.
Now, addition can be interpreted as an
operation on arrows.
Just as in the case of complex numbers.
To add vectors u and v, you place the
tail of v's arrow at the head of u's
arrow.
And now you draw a new arrow from the
tail of u to the head of u.
Now, with complex numbers, we're able to
scale them, make them bigger or smaller,
by multiplying by real numbers.
We're going to do the same thing for that
group.
So we refer to field elements such as real
numbers such as scalars.
And we use them to scale vector using
scalar vector multiplication which is
indicated in this way.
Now I use alpha, beta, gamma, Greek
letters to represent scalars in the
context of scale or vector multiplication.
The formal definition goes like this.
To multiply a vector by a scalar, you
simply multiply each of the entries by
that same scalar.
Here's an example.
Now let's interpret scaling on, as an
operation on arrows.
So here's an arrow representing the
vector 3, 1, 5.
And when we scale it by 2, we get a
vector that's in the same direction, but
is twice as long.
Scalar vector multiplication satisfiies
an algebraic property.
Multiplying a vector by beta and then
multiplying the result by alpha is the
same as multiplying the original vector
by alpha times beta.
Now, let's start with the vector 'v' say
(3,2) and consider a whole bunch of
scalar multiples using scalars ranging
from zero to one.
Each scale gives you a different size
arrow in the same direction.
So let's try that in Python.
We start with the vector v represented as
a two element list.
Now we're going to plot a whole bunch of
scaler vector multiplications.
[BLANK_AUDIO]
[SOUND] And you could see
[INAUDIBLE]
it's starting to form a line segment.
Let's do the same thing, but with a lot
more points.
What if we let the scalar alpha range
over all real numbers, rather than just
the numbers between zero and one.
Well, the scalar is bigger than one,
will give rise to arrows that are longer
than the original vector.
And the negative scalars will give rise
to arrows in the opposite direction.
Let's try it out in Python.
Our conclusion is that if you let alpha
range over all real numbers The set of
scalar multiples forms the line going
through the origin and the vector v
represented as a point.
Now, so far we know how to represent
lines and line segments that go through
the origin and a given, given vector v.
But, for purposes of say drawing maps, we
want to be able to draw line segments
that don't necessarily go through the
origin.
Well let's see how we can derive that.
We're going to use the idea of
translation.
So, we know how to draw a line segment
from zero to three two.
That's just this.
Instead of multiples Alpha times (3,2)
where alpha ranges between 0 and 1.
So here's that line segment.
What we're going to do is translate that.
Now we know the translation is performed
by doing vector addition.
So we're going to add the vector (.5,
1) to every vector in that Line segment.
And what we should get is the line
segment from 0 plus 0.5, 1 to 3, 2 plus
0.5, 1.
Scalar vector multiplication distributes
over vector addition.
[BLANK_AUDIO] For example, you can add
these two vectors and then multiply them
by 2 or you could multiply each of the
vectors by 2 and then add the results.
And you'll get the same answer.
We'll use that in coming up with another
formulation of the line segment between
two points.
So here's our current formulation of that
line segment.
We take the set of points making the line
segment from 0 to 3.2 and we add 0.5, 1.
This is not a particularry nice
formulation.
It doesn't treat the end points in a
symetric way.
So we' re going to use The distributive
law, to reformulate it.
So alpha times [3,2] plus [0.5,1] can be
rewritten in this way using the
distributive law.
Another application of the distributive
law gives us this, and we can rewrite
that in this way.
So now, we've got this expressed in terms
of the two end points and two scalar multipliers, alpha and
beta, where beta is 1 minus alpha.
So a more symmetric formulation of the
line segment between two points.
Is alpha times one point, plus beta times
the other.
Ranging over all alpha and beta that sum
to and are non-negative.
And this formulation is symmetric.
So our new formulation is alpha times one
vector, plus beta times another.
For all alpha and beta that are non
negative and sum to 1.
Expression of this form where this form,
where the scaler multipliers are non
negative and sum to one is called a
convex combination.
And what we've seen is, that the u to v
line segment, is the set of convex
combinations Of the vectors u and v.
We can use convex combinations on vectors
that represent things other than points.
Here we're going to take two vectors u
and v that are representing images.
And one example of a convex combination
assigns Alpha one half and beta one half,
this is just the average.
And the average of two images is some mix
of the two.
Now we can also use the idea of a line
segment between two faces.
So here we're taking a whole bunch of
different combinations of alpha and beta
to represent a sort of line segment
between faces.
We can get a kind of crossfade effect in
this way.
And what about the infinite line through
two given points.
We had a formulation which is not so
great, it's not symmetric between the two
points, but we can get a nicer
formulation in this way.
Alpha times one point plus beta times the
other, for all alpha and beta that sum to
one.
An expression of this form is called an
affine combination of the two vectors U
and V.
And we see that the line through U and V
consists of all affine combinations of U
and V.

Vous aimerez peut-être aussi