Definition : An inner product on a vector space 𝑉 is a function
that associates a real number < 𝒖, 𝒗 > with each pair of vectors in 𝑉 in such a way that the following axioms are satisfied for all vectors 𝒖, 𝒗, 𝒘 in 𝑉 and all scalers 𝑘. 1. < 𝒖, 𝒗 > =< 𝒗, 𝒖 > [Symmetry axiom] 2. < 𝒖 + 𝒗, 𝒘 > =< 𝒖, 𝒘 >+< 𝒗, 𝒘 > [Additive axiom]
3. < 𝑘𝒖, 𝒗 > = 𝑘 < 𝒖, 𝒗 >
4. < 𝒗, 𝒗 > ≥ 0 and < 𝒗, 𝒗 > = 0 if and only if 𝒗 = 𝟎 [Positivity axiom] Inner Product Spaces
Definition : If 𝑉 = 𝑅𝑛 and inner product define on 𝑉 as
< 𝒖, 𝒗 > = 𝒖. 𝒗 = 𝑢1 𝑣1 + 𝑢2 𝑣2 +. . . . 𝑢𝑛 𝑣𝑛 This inner product is called Euclidian inner product (or Standard inner product) on 𝑅𝑛 . We call 𝑅𝑛 with Euclidian inner product Euclidian n-space. Inner Product Spaces
Definition: If 𝑉 is a inner product space, then the norm
(or length) of a vector 𝒗 in 𝑉 is denoted by 𝒗 and defined by 𝒗 = < 𝒗, 𝒗 > A vector of norm 1 is called a unit vector. and distance between two vectors is denoted by 𝑑(𝒖, 𝒗) and is defined by 𝑑 𝒖, 𝒗 = 𝒖 − 𝒗 = < 𝒖 − 𝒗, 𝒖 − 𝒗 > Inner Product Spaces Theorem: If 𝒖 and 𝒗 are vectors in a inner product space 𝑉, and if 𝑘 is any scaler, then a) 𝒗 ≥ 0 and 𝒗 = 0 if and only if 𝒗 = 𝟎. b) 𝑑 𝒖, 𝒗 ≥ 0 and 𝑑 𝒖, 𝒗 = 0 if and only if 𝒖 = 𝒗. c) 𝑑 𝒖, 𝒗 = 𝑑 𝒗, 𝒖 . d) 𝑘𝒗 = 𝑘 𝒗 . Inner Product Spaces Definition: If 𝑤1 , 𝑤2 , . . . . . , 𝑤𝑛 are positive real numbers. 𝒖 = (𝑢1 , 𝑢2 , . . . , 𝑢𝑛 ) and 𝒗 = (𝑣1 , 𝑣2 , . . . , 𝑣𝑛 ) are vectors in 𝑅𝑛 ,then the formula < 𝒖, 𝒗 > = 𝑤1 𝑢1 𝑣1 + 𝑤2 𝑢2 𝑣2 +. . . . 𝑤𝑛 𝑢𝑛 𝑣𝑛 defines an inner product on 𝑅 𝑛 and we call the weighted Euclidian inner product with weights 𝑤1 , 𝑤2 , . . . . . , 𝑤𝑛 . Inner Product Spaces Example 1: (Weighted Euclidian Inner Product) Let 𝒖 = (𝑢1 , 𝑢2 ) and 𝒗 = (𝑣1 , 𝑣2 ) be vectors in 𝑅2 . Is the weighted Euclidian inner product defined by < 𝒖, 𝒗 > = 3𝑢1 𝑣1 + 2𝑢2 𝑣2 Satisfies the four inner product axioms. Inner Product Spaces Example 2: (norm and distance) It is important to keep in mind that norm and distance depends on inner product being used. Let 𝒖 = (1,0) and 𝒗 = (0,1) in 𝑅2 . With Euclidian inner product: 𝒖 = 12 + 02 = 1 and 𝑑 𝒖, 𝒗 = 𝒖 − 𝒗 = (1, −1) = 12 + (−1)2 = 2 With Weighted Euclidian inner product: < 𝒖, 𝒗 > = 3𝑢1 𝑣1 + 2𝑢2 𝑣2 𝒖 = < 𝒖, 𝒖 >= 3 1 1 + 2(0)(0) = 3 𝑑 𝒖, 𝒗 = 𝒖 − 𝒗 = < 1, −1 , (1, −1) > = 3 1 1 + 2 −1 −1 = 5 Inner Product Spaces Example: Let 𝑉 = 𝑅𝑛 . Let 𝒖 and 𝒗 are vectors in 𝑅𝑛 that are expressed in column vectors 𝑛 × 1, in such a case the formula < 𝒖, 𝒗 > = 𝒖𝑇 𝒗 defines an inner product on 𝑅 𝑛 . Example: (An inner product on 𝑴𝒏𝒏 ) If 𝑈 and 𝑉 are 𝑛 × 𝑛 matrices then the formula < 𝑈, 𝑉 > = 𝑡𝑟(𝑈 𝑇 𝑉) Defines an inner product on 𝑀𝑛𝑛 . < 𝑈, 𝑉 > = 𝑡𝑟 𝑈 𝑇 𝑉 = 𝑢1 𝑣1 + 𝑢2 𝑣2 + 𝑢3 𝑣3 + 𝑢4 𝑣4 Inner Product Spaces Example: (The standard inner product on 𝑷𝒏 ) If 𝑝 = 𝑎0 + 𝑎1 𝑥+ . . . +𝑎𝑛 𝑥 𝑛 and q = 𝑏0 + 𝑏1 𝑥+ . . . +𝑏𝑛 𝑥 𝑛 are polynomials in 𝑃𝑛 , then < 𝑝, 𝑞 > = 𝑎0 𝑏0 + 𝑎1 𝑏1 +. . . . . . +𝑎𝑛 𝑏𝑛 defines an inner product on 𝑃𝑛 called standard inner product. And norm is 𝑝 = < 𝑝, 𝑝 >= 𝑎0 2 + 𝑎1 2 + . . . . +𝑎𝑛 2 Example: (Another inner product on 𝑷𝒏 ) < 𝑝, 𝑞 > = 𝑝 𝑥0 𝑞 𝑥0 + 𝑝 𝑥1 𝑞 𝑥1 +. . . . +𝑝 𝑥𝑛 𝑞 𝑥𝑛 Defines an inner product on 𝑷𝒏 . Where 𝑥0 , 𝑥1 , . . , 𝑥𝑛 are distinct real numbers. And norm of polynomial is 𝑝 = < 𝑝, 𝑝 >= [𝑝 𝑥0 ]2 +[𝑝 𝑥1 ]2 + . . . +[𝑝 𝑥𝑛 ]2 Inner Product Spaces Example: Let 𝑉 = 𝑃2 and 𝑥0 = −2, 𝑥1 = 0, 𝑥2 = 2 Compute < 𝑝, 𝑞 > and 𝑝 for polynomials 𝑝 = 𝑝 𝑥 = 𝑥 2 and 𝑞 = 𝑞 𝑥 = 1 + 𝑥. Angle and orthogonality in inner product spaces Theorem: (Cauchy-Schwarz inequality) If 𝒖 and 𝒗 are vectors in a vector space 𝑉, then | < 𝒖, 𝒗 > | ≤ 𝒖 𝒗 Remark: Angle 𝜃 between 𝒖 and 𝒗 defines as −1 < 𝒖, 𝒗 > 𝜃 = 𝑐𝑜𝑠 𝒖 𝒗 Example: Let 𝑉 = 𝑅4 have Euclidian inner product. Find the angle 𝜃 between the vectors 𝒖 = 4,3,1, −2 , 𝒗 = (−2,1,2,3) Angle and orthogonality in inner product spaces Theorem: If 𝒖, 𝒗, 𝒘 are vectors in a vector space 𝑉, then a) 𝒖 + 𝒗 ≤ 𝒖 + 𝒗 [Triangle inequality for vectors] b) 𝑑 𝒖, 𝒗 ≤ 𝑑 𝒖, 𝒘 + 𝑑(𝒘, 𝒗) [Triangle inequality for distance]
Definition: Two vectors 𝒖 and 𝒗 are in an inner product space
are called orthogonal if < 𝒖, 𝒗 > = 0 Angle and orthogonality in inner product spaces Example: (Orthogonality depends on inner product) The vectors 𝒖 = (1,1) and 𝒗 = (1, −1) are orthogonal with respect to Euclidian inner product on 𝑅2 , since < 𝒖, 𝒗 > = 𝒖. 𝒗 = (1)(1) + (1)(−1) = 0 But they are not orthogonal with respect to weighted Euclidian inner product < 𝒖, 𝒗 > = 3𝑢1 𝑣1 + 2𝑢2 𝑣2 , since < 𝒖, 𝒗 > = 3 1 1 + 2 1 −1 = 1 ≠ 0 Angle and orthogonality in inner product spaces Definition: If 𝑊 is a subspace of a vector space 𝑉, then the set of all vectors in 𝑉 that are orthogonal to every vector in 𝑊 is called the orthogonal compliment of 𝑾 and is denoted by 𝑊 ⊥ .
Theorem: If 𝑊 is a subspace of inner product space 𝑉, then
1. 𝑊 ⊥ is a subspace of 𝑉 2. 𝑊 ∩ 𝑊 ⊥ = {𝟎}. Angle and orthogonality in inner product spaces Example: Basis for orthogonal compliment Let 𝑊 be the subspace of 𝑅6 spanned by the vectors 𝒘𝟏 = 1,3, −2,0,2,0 , 𝒘𝟐 = 2,6, −5, −2,4, −3 𝒘𝟑 = 0,0,5,10,0,15 , 𝒘𝟒 = (2,6,0,8,4,18) Find a basis for 𝑊 ⊥ (orthogonal compliment).