MA212 Further Mathematical Methods Lecture 1: Some revision of MA100 linear algebra Dr James Ward ■ Vector spaces and subspaces ■ Linear independence and span The London School of Economics and Political Science Department of Mathematics Lecture 1, page 1 Information ■ Exercises 11 is on Moodle. ▶ Attempt questions 2, 3, 5, 6 and 9. ▶ Revision of MA100 linear algebra you should know. ▶ Submit your written solutions on Moodle by Thursday or follow any instructions laid down by your class teacher. ■ Extra Examples Sessions ▶ Start in Week 2, Friday 13:00–14:00 in OT. ▶ Reply to the post in the Moodle forum if you want me to cover anything in particular. ■ Classes start in Week 2. See your timetable for details. Lecture 1, page 2 Vector spaces A real vector space is a non-empty set, V , whose elements are called vectors that is closed under two algebraic operations, namely Closure under vector addition: Given the vectors u, v ∈ V , we can add them to form the vector u + v ∈ V ; Closure under scalar multiplication: Given the vector u ∈ V and the scalar α ∈ R, we can multiply the vector by the scalar to form the vector αu ∈ V . In MA100, the scalars were always real numbers giving us a real vector space. Later in the course, we will see what happens when the scalars are complex numbers. Lecture 1, page 3 Vector spaces In each vector space V , there is a special element called the zero vector and we denote this by 0. This has the property that Zero vector: For all v ∈ V , v + 0 = v . There is also a special scalar, called the unit scalar and we denote this by 1. This has the property that Unit scalar: For all v ∈ V , 1v = v . Lecture 1, page 4 Vector spaces Also, every vector in the vector space has an inverse under the operation of vector addition Inverse vectors: For all v ∈ V , there is a vector −v ∈ V such that v + (−v ) = 0. It should come as no surprise that the vector −v is the vector v multiplied by the scalar −1, i.e. −v = (−1)v . Lecture 1, page 5 Vector spaces Lastly, the operations of vector addition and scalar multiplication must satisfy some obvious properties. For all vectors u, v , w ∈ V and all scalars α, β ∈ R, ■ u + v = v + u, ■ u + (v + w ) = (u + v ) + w , ■ α(u + v ) = αu + αv , ■ (α + β)u = αu + βu, ■ α(βu) = (αβu). And these should all make sense given what you know about how vectors and scalars behave. Lecture 1, page 6 Example The set a 1 3 R = a2 a1 , a2 , a3 ∈ R , a3 where vector addition and scalar multiplication are defined by a1 b1 a1 + b1 a1 αa1 a2 + b2 = a2 + b2 and α a2 = αa2 , a3 b3 a3 + b3 a3 αa3 is a real vector space. 0 Observe that the zero vector in this space is 0 = 0. 0 Lecture 1, page 7 Example The set ( a1 a2 a3 a4 ) ! a1 , a2 , a3 , a4 ∈ R , where vector addition and scalar multiplication are defined by ! ! ! a1 a2 b1 b2 a1 + b1 a2 + b2 + = a3 a4 b3 b4 a3 + b3 a4 + b4 and a1 a2 α a3 a4 ! = αa1 αa2 αa3 αa4 ! , is a real vector space. Observe that the zero vector in this space is 0 = ! 0 0 . 0 0 Lecture 1, page 8 Example The set of all functions f : [0, 1] → R with the usual ‘point-wise’ operations of vector addition and scalar multiplication is a real vector space. We’ll call this vector space F[0, 1]. Observe that the zero vector in this space is the function 0 : [0, 1] → R given by 0(x) = 0 for all x ∈ [0, 1]. Note: By ‘point-wise’ operations we just mean that, given the functions f , g : [0, 1] → R and the scalar α ∈ R, we have the functions f + g and αf whose values are given by (f + g )(x) = f (x) + g (x) and (αf )(x) = αf (x), for all x ∈ [0, 1]. Lecture 1, page 9 Subspaces Suppose that V is a vector space and U is a subset of V . If U is also a vector space, then we call U a subspace of V . Essentially, this means that U is a subspace of V if it is a non-empty subset of V that is closed under the operations of vector addition and scalar multiplication being used in V . Theorem: U is a subspace of a real vector space V if and only if U is a non-empty subset of V which is both CUVA: For all u, v ∈ U, u + v ∈ U and CUSM: For all u ∈ U and α ∈ R, αu ∈ U. Lecture 1, page 10 Example Show that the set x U = y 2x − y + z = 0 z is a subspace of R3 . Lecture 1, page 11 Example Show that the set of all continuous functions f : [0, 1] → R is a subspace of F[0, 1]. Lecture 1, page 12 Linear independence Definition: Suppose that V is a vector space. A finite set of vectors, {v1 , v2 , . . . , vk } ⊆ V is linearly independent if the only solution to the equation α1 v1 + α2 v2 + · · · + αk vk = 0 is α1 = α2 = · · · = αk = 0. Note: We often call this the trivial solution to the equation as it is always one of the solutions. The key for linear independence is that it is the only solution! Lecture 1, page 13 Example Show that the set of vectors 2 0 1 3 , 2 , 0 −1 1 1 is linearly dependent. Lecture 1, page 14 Example Show that the set of vectors 0 0 1 3 , 2 , 0 1 1 1 is linearly independent. Lecture 1, page 15 A test for linear independence Question: How can we easily decide if a finite set of vectors in Rn is linearly independent? Answer: Write the vectors as the rows of a matrix and perform row operations to get the row-echelon form (REF) of the matrix. The vectors are linearly independent if and only if the REF does not have a row of zeroes. Recall: We have three types of row operation: (1) exchange two rows, (2) multiply a row by a non-zero scalar, (3) add a non-zero scalar multiple of one row to another row. The REF has a one as the first entry of each non-zero row and each such ‘leading one’ only has zeroes below it. Lecture 1, page 16 Example Is the set of vectors −1 2 1 1 , 0 , 3 1 −1 0 linearly independent? Lecture 1, page 17 Example Is the set of vectors 2 0 1 3 , 2 , 0 −1 1 1 linearly independent? Lecture 1, page 18 More tests for linear independence When we are dealing with a set of n vectors in Rn , the following theorem from MA100 gives us a number of useful ways to test for linear independence. Theorem: For a square matrix A, the following statements are equivalent. (1) The rows of A are linearly independent. (2) The REF of A does not have a row of zeroes. (3) The determinant of A is non-zero. (4) The inverse of A exists. Lecture 1, page 19 Other thoughts on linear independence When we are just dealing with two vectors, we also have the scalar multiple test for linear independence. Theorem: Suppose that V is a vector space. Two non-zero vectors u, v ∈ V are linearly independent if and only if there is no scalar α such that u = αv . Also remember that any set that contains the zero vector is linearly dependent. Theorem: Suppose that V is a vector space and U ⊆ V . If 0 ∈ U, then U is linearly dependent. Lecture 1, page 20 Linear combinations Definition: Suppose that V is a vector space and that S ⊆ V is the set of vectors S = {u1 , u2 , . . . , un }. Any vector of the form α1 u1 + α2 u2 + · · · + αn un for scalars α1 , α2 , . . . , αn is a linear combination of the vectors in S. As V is closed under vector addition and scalar multiplication, we know that every such linear combination will also be a vector in V . Lecture 1, page 21 Linear span Indeed, if we were to take the set of all possible linear combinations of the vectors in S, we would get the linear span of S. Definition: Suppose that V is a vector space and that S ⊆ V is the set of vectors S = {u1 , u2 , . . . , un }. The linear span of S is the set of vectors Lin(S) = {α1 u1 +α2 u2 +· · ·+αn un | α1 , α2 , . . . , αn are scalars}. Instead of ‘Lin(S)’, you may also see ‘span(S)’ being used. Lecture 1, page 22 Linear spans give subspaces The linear span of a set of vectors is one way of generating subspaces. Theorem: If V is a vector space and S ⊆ V , then Lin(S) is a subspace of V . Indeed, we can even show that the linear span of S ⊆ V is the smallest subspace of V that contains all of the vectors in S. Theorem: Suppose that V is a vector space and S ⊆ V . If U is any subspace of V with S ⊆ U, then Lin(S) ⊆ U. That is, any subspace of V that contains all of the vectors in S will also contain all of the vectors in Lin(S). Lecture 1, page 23 Example In R3 , consider the linear span 1 1 Lin 1 = α 1 α ∈ R . 1 1 It is a subspace of R3 and it is the smallest subspace of R3 that contains the vector (1, 1, 1)t . Geometrically, in R3 , we can think of this as a line through the origin in the direction (1, 1, 1)t . Lecture 1, page 24 Example In F[0, 1], consider the functions fn : [0, 1] → R given by fn (x) = x n for n = 0, 1, 2, . . .. The linear span Lin{f0 , f1 , . . . , fn } = {α0 f0 +α1 f1 +· · ·+αn fn | α0 , α1 , . . . , αn ∈ R} is a subspace of F[0, 1] and it is the smallest subspace of F[0, 1] that contains all of the functions f0 , f1 , . . . , fn . For all x ∈ [0, 1], these functions have values given by (α0 f0 + α1 f1 + · · · + αn fn )(x) = α0 + α1 x + · · · + αn x n , and so we can think of this linear span as the set of all polynomials of degree at most n. We’ll call this vector space Pn [0, 1]. Lecture 1, page 25 Linear independence and linear span Lastly, we can use linear spans to say some useful things about linear independence. Theorem: Suppose that V is a vector space and that S ⊆ V is a finite set of vectors. The following statements are equivalent. (1) S is linearly independent. (2) Each vector in Lin(S) can be written as a linear combination of the vectors in S in exactly one way. (3) For each u in S, u is not in Lin(S \ {u}). Note: (3) says that each vector u ∈ S can not be written as a linear combination of the vectors left in S after you remove u. Lecture 1, page 26