Lesson 2: Vector Spaces September 27, 2014 Lesson 2: Vector Spaces Vector Spaces Let V be a set, and let +, · be two operations, the first one (sum) defined between the elements of V , and the second one (product by scalars) defined between V , and the elements of R (resp. C). We say that (V , +, ·) is a vector space over R (resp. C) if the following properties hold: (i) (V , +) is a conmutative group: Internal Law: ∀~u , ~v ∈ V , ~u + ~v ∈ V . ~ ∈ V , ~u + (~v + w ~ ) = (~u + ~v ) + w ~. Associative: ∀~u , ~v , w Neutral element: there exists ~0 ∈ V such that ~u + ~0 = ~u . Inverse element: for all ~u ∈ V , there exists −~u ∈ V such that ~u + (−~u ) = ~0. Conmutative: ∀~u , ~v ∈ V , ~u + ~v = ~v + ~u . Lesson 2: Vector Spaces Vector Spaces Let V be a set, and let +, · be two operations, the first one (sum) defined between the elements of V , and the second one (product by scalars) defined between V , and the elements of R (resp. C). We say that (V , +, ·) is a vector space over R (resp. C) if the following properties hold: (ii) The operation · satisfies that: ∀λ ∈ R (resp. C), ∀~u , ~v ∈ V , λ · (~u + ~v ) = λ~u + λ~v . ∀λ, µ ∈ R (resp. C), ∀~u , ~v ∈ V , (λ + µ) · ~u = λ~u + λ~v . ∀λ, µ ∈ R (resp. C), ∀~u , ~v ∈ V , λ · (µ~u ) = (λ · µ) · ~u . 1 · ~u = ~u . Lesson 2: Vector Spaces Vector Spaces In order to make explicit whether we work over R or C, one writes (V (R), +, ·) or (V (C), +, ·). In the sequel, we will assume that we work over R, although the results are equivalent for C. In fact, vector spaces can be defined over sets other than R or C (fields). Examples: Khan Academy (click) Lesson 2: Vector Spaces Vector Spaces Proposition Let (V (R), +, ·) be a vector space over R. Then for all λ ∈ R and for all ~u ∈ V , it holds that: (1) λ · ~0 = ~0. (2) 0 · ~u = ~0. (3) λ · ~u = ~0 if and only if λ = 0 or ~u = ~0. (4) λ · (−~u ) = (−λ) · ~u = −(λ · ~u ). Lesson 2: Vector Spaces Linear Dependence. Bases. Definition A linear combination of vectors ~u1 , ~u2 , . . . , ~un is another vector of the form λ1~u1 + λ2~u2 + · · · + λn~un where λi ∈ R for i = 1, . . . , n. Remark The vector ~0 can be considered as a linear combination of any vectors ~u1 , ~u2 , . . . , ~un , because ~0 = 0 · ~u1 + 0 · ~u2 + · · · + 0 · ~un . Lesson 2: Vector Spaces Linear Dependence. Bases. Definition We say that {~u1 , ~u2 , . . . , ~un } are linearly dependent (l.d.) if at least of them is a linear combination of the rest. If {~u1 , ~u2 , . . . , ~un } are not l.d., we say that they are linearly independent (l.i.). Lesson 2: Vector Spaces Linear Dependence. Bases. For vectors in Rn : in order to check the linear independence of several vectors, form a matrix with them as columns (or rows), and compute the rank (Example: Khan Academy (click)) The above idea can be extended to other cases (but we need the notion of coordinates of a vector, to do this; we will see it later). Lesson 2: Vector Spaces Linear Dependence. Bases. Theorem The vectors {~u1 , ~u2 , . . . , ~un } are linearly independent if and only the only linear combination of them fulfilling λ1~u1 + · · · + λ~un = ~0 satisfies λ1 = · · · = λn = 0. Proof.: Khan Academy (click) Lesson 2: Vector Spaces Linear Dependence. Bases. Definition We say that S = {~u1 , . . . , ~un } is a spanning set of V if any vector in V can be written as a linear combination of the vectors in S. Definition We say that B = {~u1 , . . . , ~un } is a basis of V if it is a spanning set of V and they are linearly independent. Examples: Khan Academy (click) Important remark: vector spaces may have infinitely many bases!! Lesson 2: Vector Spaces Linear Dependence. Bases. Definition We say that a vector space has finite dimension if it has a basis consisting of finitely many vectors. Examples: 1 Rn has dimension n, since it admits the basis (called the canonical basis) {(1, 0, 0, . . . , 0), (0, 1, 0, . . . , 0), . . . , (0, 0, 0, . . . , 1)} 2 3 Notation: ~e1 = (1, 0, . . . , 0), ~e2 = (0, 1, . . . , 0), . . ., ~en = (0, 0, . . . , 1). Matrices of fixed dimension, polynomials of fixed degree. At least an example of a vector space which is NOT this kind? Lesson 2: Vector Spaces Linear Dependence. Bases. Theorem If V has finite dimension, then all the bases of V have the same number of vectors (the dimension of V , dim(V )). Intuitively, the dimension of a vector space is the number of parameters that one needs to specify in order to identify a concrete element in the space. Lesson 2: Vector Spaces Linear Dependence. Bases. Theorem Let V be a vector space with finite dimension, and let dim(V ) = n. Then the following statements are true: (i) If S spans V , then you can extract a basis from S. (ii) Every system consisting of more than n vectors is linearly dependent. (iii) Every spanning system contains at least n vectors. (iv) Given B = {~u1 , . . . , ~un } ⊂ V (a subset of exactly n vectors), the following statements are equivalent: (a) B is a basis; (b) B is linearly independent; (iii) B is a spanning system. Lesson 2: Vector Spaces Linear Dependence. Bases. Definition Let B = {~u1 , . . . , ~un } be a basis of V , and let ~v ∈ V . The coordinates of ~v with respect to B are the scalars λ1 , . . . , λn ∈ R such that ~v = λ1~u1 + . . . + λn~un Usually we write ~v = (λ1 , . . . , λn )B ; we also say ~v has coordinates (λ1 , . . . , λn ) in B. Lesson 2: Vector Spaces Linear Dependence. Bases. Theorem Let V be a vector space of finite dimension n, and let B = {~u1 , . . . , ~un } be a basis of V . Then every vector ~v ∈ V has unique coordinates with respect to B. Proof. + Examples: Khan Academy (click) Lesson 2: Vector Spaces Vector Subspaces. Definition Let (V (R), +, ·) be a vector space. We say that W ⊂ V is a vector subspace of V if (W (R), +, ·) has also a structure of vector space. Lesson 2: Vector Spaces Vector Subspaces. Theorem W ⊂ V is a vector subspace if and only if ∀ ~u , ~v ∈ W , ∀λ, µ ∈ R, λ~u + µ~v ∈ W (i.e. if and only if every linear combination of two vectors in W , stays in W ). Examples: Khan Academy (click) Observation: If W is a vector subspace, then ~0 ∈ W . Lesson 2: Vector Spaces Vector Subspaces. Since a vector subspace is in fact a vector space inside of another bigger vector space, it makes sense to speak about its dimension, about bases, etc. Again, the dimension of a vector subspace is the number of parameters defining an element of the subspace. Examples: Khan Academy (click) Lesson 2: Vector Spaces Vector Subspaces. Definition let S = {~u1 , . . . , ~un } be a subset of V . The linear variety spanned by S (or simply the linear span of S) is the set consisting of all the vectors which are linear combinations of the vectors in S, i.e. L(S) = {~x ∈ V |~x = λ1~u1 + · · · + λn~un } L(S) is a vector subspace, and its dimension is the rank of the system S, i.e. the number of linearly independent vectors in S. In R2 , the linear span of a vector is a line. In R3 , the linear span of one vector is also a line; the linear span of two independent vectors is a plane. In Rn , n ≥ 4, one vector spans a line, two independent vectors span a plane, three independent vectors span a space,... Lesson 2: Vector Spaces Vector Subspaces. Definition Let S1 , S2 be two vector subspaces of V . Then we define: (i) The intersection S1 ∩ S2 of S1 , S2 : S1 ∩ S2 = {~x ∈ V |~x ∈ S1 and ~x ∈ S2 } (ii) The sum S1 + S2 : S1 + S2 = {~x ∈ V |~x = ~u + ~v , ~u ∈ S1 , ~v ∈ S2 } Lesson 2: Vector Spaces Vector Subspaces. Theorem S1 ∩ S2 and S1 + S2 are vector subspaces. Moreover, it holds that dim(S1 + S2 ) = dim(S1 ) + dim(S2 ) − dim(S1 ∩ S2 ) Question: Let S1 , S2 be two different planes in R3 containing ~0 and sharing a same line. What is S1 + S2 ? Lesson 2: Vector Spaces Vector Subspaces. Observations: S1 ∩ S2 always contains ~0; it can happen S1 ∩ S2 = {~0}. If B1 = {~u1 , . . . , ~um } is a basis of S1 , and B2 = {~u1′ , . . . , ~up′ } is a basis of S2 , then S1 + S2 = L(~u1 , . . . , ~um , ~u1′ , . . . , ~up′ ) Furthermore, a basis of S1 + S2 can be found by taking the linearly independent vectors in ~u1 , . . . , ~um , ~u1′ , . . . , ~up′ . Lesson 2: Vector Spaces Vector Subspaces. Definition Two subspaces S1 , S2 are said to be independent if every vector ~x ∈ S1 + S2 can be expressed in a unique way as a sum of vectors of S1 and S2 , i.e. if the decomposition ~x = ~u + ~v where ~u ∈ S1 and ~v ∈ S2 , is unique. In this L situation, one also says that the sum is direct, which is denoted S1 S2 . Theorem L S1 S2 if and only if S1 ∩ S2 = {~0}. Proof. + Example (Khan Academy (click) Lesson 2: Vector Spaces Vector Subspaces. Equations of a vector subspace: Khan Academy (click) 1 2 3 Vector equation. Parametric equations. Implicit equations. Change of basis: Khan Academy (click) Lesson 2: Vector Spaces