Linear Algebra Differential Equations Math 54 Lec 005 (Dis 501) July 10, 2014

Transcription

1 Vector space R n A vector space R n is the set of all possible ordered pairs of n real numbers So, R n = {(a, a,, a n ) : a, a,, a n R} a a We abuse the notation (a, a,, a n ) instead of sometimes a n The meaning of A being a subset of B is that for every element a in A it is included in B So, R is not a subset of R However, R {} is a subset of R Since what we mean by R {} is R {} = {(a, b, ) : a, b R} A subspace of R n What does it mean by that H is a subspace of R n? First of all, of course, we require H to be a subset of R n Hence, for instance, R can never be a subspace of R However, R {} can be a subspace of R and it is indeed Now, for a subset H to be a subspace we require three conditions : ) H has the zero vector ) H is closed under addition ) H is closed under scalar multiplication ) In the case of R, the zero vector of R is (, ) Hence, for H to be a subspace of R, H should have (, ) If the case is R, then H should have (,, ) For the next two conditions, let s think about the meaning of closed, ) Example The set of all odd numbers Z odd and the set of all even numbers Z even Z odd is NOT closed under addition, meanwhile Z even is closed under addition Why? In this sense for H to be a subspace of R n, we require addition of any two vectors in H should be inside of H again Also, we require the multiplication of any vector in H with any real number belongs to H If these conditions are satisfied H is a subspace of R n

2 A vector space P n The set of all polynomials of degree at most n will be denoted by P n Examples of a polynomial of degree 5 could be So, these polynomials below is also inside of P 5 : t 5 t 4 + t t + t 5 t + t t 5 + t 5 t 4 t + t + t + t + Here, (one) is a polynomial of degree P n is also a vector space What a vector means in this specific vector space is a polynomial of degree at most n The zero vector of the vector space P n is + t + + t n, or shortly Please note that we do not define the multiplication between two vectors So, when we think of P n as a vector space, we do NOT define t t or (t + ) (t 4 t + ) We only have the addition of two vectors (in other words, two polynomials) and the scalar multiplication of one vector (in other words, one polynomial) and one scalar (in other words, a real number) The addition of two vectors t 5 + t t + t + 5t + t is t 5 + t + 5t + t + The scalar multiplication of a vector t 4 t + t + 8 and a scalar is t 4 + t 4t 56 A subspace of P n Similarly as before, for a set H to be a subspace of P n, we first require H to be a subset of P n And then, it is the same as before (the case of R n ), we require three conditions ) H should have the zero vector ) H is closed under addition ) H is closed under scalar multiplication Is a subspace of P 4? S = {p(t) P 4 : p() = } Is a subspace of P 4? S = {p(t) P 4 : p() = }

3 Col A and Nul A Let A be an m n matrix Col A is the column space of a matrix A Given a matrix A, it is NOT a matrix or NOT a vector but a set of specific vectors Col A = {Au : u R n } = Span{column vectors of A} = {all linear combinations of column vectors of A} For example, given a matrix A = the column space Col A contains vectors = + + and also = + ( ) + Obviously, Col A contains the zero vector in R 4, = + + Here, please note that, by the definition of the multiplcation of an m n matrix and a vector in R n = + + = Hence, we can say that Col A = {Au : u R n } in general Conventionally, we use notations e, e,, e n for e =, e =,, e n = Using these vectors, we easily know that Ae is the first column of A and Ae i is the ith column of A So, in other words, we can say Col A =Span{Ae, Ae,, Ae n } Even though we need to specify where those e, e,, e n belong to, we just abuse the notations

4 Elementary row operations We will skip the part that every elementary row operation has a particular relation with an elementary matrix, that is, Doing an elementary row operation to A = Converting A into EA where E is the corresponding matrix As we have seen before, Ae i represents the ith column of A Hence, if we have a relation like Ae + 4Ae Ae + Ae n =,then we also have EAe + 4EAe EAe + EAe n = E(Ae + 4Ae Ae + Ae n ) = E = From this example, we can figure out that elementary row operations do not change linear dependence relation of column vectors However, it might change Col A In other words, there is no reason for Col A = Col EA We also have some counterexamples : The column space of the former matrix contains contain that vector obviously, but the column space of the latter matrix does not As we have discussed this thing many times, 4

5 Nul A We define Nul A as the set Nul A = {u R n : Au = Conventionally, we use the notation or for the zero vector } Is Nul A a subspace of R n? The answer is yes Because First of all, Nul A is obviously a subset of R n ) The zero vector : Since A =, we can get Nul A ) Closed under addition : Suppose that u and v is in Nul A In other words, we are assuming that Au = and Av = We want to check whether or not u+v is in Nul A However, A(u+v) = Au+Av = + = Hence, u+v Nul A ) Closed under scalar multiplication : As we have proven that ), ), and ) are all true, Nul A is a subspace of R n Could Nul A and Col A be the same? Yes, it sometimes happens We have an example : {( has its Col A as Span A = ( )} {( At the same time, Nul A is Span However, in general, Col A R m and Nul A R n, so they couldn t be the same unless m = n since they belong to different vector spaces unless m = n ) )} The Rank Theorem Let A be an m n matrix Then, dim Nul A + dim Col A = n In other words, we can change dim Col A into rank A Prove it! 5

6 Linear Transformation As we have gone through several times, what it means by for a map or a function T being a linear transformation from a vector space V to a vector space W is ) T (v + v ) = T (v ) + T (v ) for all v, v V ) T (cv) = ct (v) for all v V and c R There is one fact that we can easily get from the definition : set c = in ), then we get T () = We have learned one good example of a linear transformation : Matrix multiplication For example, let s think about a matrix ( ) 5 A = As we multiply a vector in R, we get a vector in R like this ; ( ) ( ) A =, A =, A 5 6 ( ) 8 = So this matrix A gives a function from R to R conditions For this function to be a linear transformation, we require two ) A(u + v) = Au + Av for all u, v R ) Acu = cau for all u R and c R Those two conditions are trivially true in this case In this sense, every m n matrix gives a linear transformation from R n to R m Here, let s recall Col A and Nul A Col A = {Au : u R n } & Nul A = {u R n : Au = } Similarly, for a linear transformation T from V to W, we define im T and ker T as following : im T = {T (v) : v V } & ker T = {v V : T (v) = } Then, im T = Col A and ker T =Nul A when T is the linear transformation defined by the matrix multiplication of A Remember that Col A is a subspace of R m and Nul A is a subspace of R n Similarly, im T is a subspace of W and ker T is a subspace of V How could you figure out that fact? Note that R and R are vector spaces 6

7 Let s take a look at another example Let T be the map from P to R defined as T (p(t)) = p() Note that P is a vector space and R is also a vector space Hence, T is a map from a vector space to a vector space In fact, T : P R is a linear transformation In order to show that we need two conditions ) For every v and v P, T (v + v ) = T (v ) + T (v )? ) For every v P and c R, T (cv) = ct (v)? Another way to show that S = {p(t) P 4 : p() = } is a subspace of P 4