🧚🏽‍♀️Abstract Linear Algebra I Unit 9 – Gram-Schmidt Orthogonalization

Key Concepts and Definitions

• Orthogonality: Two vectors $\mathbf{u}$ and $\mathbf{v}$ are orthogonal if their inner product $\langle \mathbf{u}, \mathbf{v} \rangle = 0$
• Orthonormal set: A set of vectors $\{\mathbf{v}_1, \mathbf{v}_2, \ldots, \mathbf{v}_n\}$ is orthonormal if each vector has unit length and any two distinct vectors are orthogonal
  • Mathematically, $\langle \mathbf{v}_i, \mathbf{v}_j \rangle = \delta_{ij}$, where $\delta_{ij}$ is the Kronecker delta
• Orthonormal basis: An orthonormal set that spans the entire vector space
• Inner product: A generalization of the dot product that measures the similarity between two vectors
  • Denoted as $\langle \mathbf{u}, \mathbf{v} \rangle$ for vectors $\mathbf{u}$ and $\mathbf{v}$
• Projection: The process of decomposing a vector into components parallel and perpendicular to a given direction
  • The projection of $\mathbf{u}$ onto $\mathbf{v}$ is given by $\text{proj}_{\mathbf{v}}\mathbf{u} = \frac{\langle \mathbf{u}, \mathbf{v} \rangle}{\langle \mathbf{v}, \mathbf{v} \rangle}\mathbf{v}$
• Linear independence: A set of vectors is linearly independent if no vector can be expressed as a linear combination of the others
• Span: The set of all linear combinations of a given set of vectors

Historical Context and Applications

• Gram-Schmidt process developed by Jørgen Pedersen Gram and Erhard Schmidt in the early 20th century
• Widely used in various fields of mathematics, physics, and engineering
• Signal processing: Orthogonal signals minimize interference and simplify analysis
  • Orthogonal frequency-division multiplexing (OFDM) in wireless communications
• Quantum mechanics: Orthonormal bases represent quantum states and simplify calculations
  • Eigenstates of observables form orthonormal bases in Hilbert spaces
• Numerical linear algebra: Orthogonalization improves stability and accuracy of algorithms
  • Least squares problems and matrix decompositions (QR decomposition)
• Computer graphics: Orthonormal bases simplify transformations and lighting calculations
• Cryptography: Orthogonal matrices used in designing secure encryption schemes

The Gram-Schmidt Process: Step-by-Step

1. Start with a linearly independent set of vectors $\{\mathbf{v}_1, \mathbf{v}_2, \ldots, \mathbf{v}_n\}$
2. Set $\mathbf{u}_1 = \mathbf{v}_1$ and normalize it: $\mathbf{e}_1 = \frac{\mathbf{u}_1}{\|\mathbf{u}_1\|}$
3. For $i = 2, 3, \ldots, n$:
   • Compute the projection of $\mathbf{v}_i$ onto each previous orthonormal vector: $\text{proj}_{\mathbf{e}_j}\mathbf{v}_i = \frac{\langle \mathbf{v}_i, \mathbf{e}_j \rangle}{\langle \mathbf{e}_j, \mathbf{e}_j \rangle}\mathbf{e}_j$ for $j = 1, 2, \ldots, i-1$
   • Subtract the projections from $\mathbf{v}_i$ to obtain $\mathbf{u}_i$: $\mathbf{u}_i = \mathbf{v}_i - \sum_{j=1}^{i-1}\text{proj}_{\mathbf{e}_j}\mathbf{v}_i$
   • Normalize $\mathbf{u}_i$ to get the next orthonormal vector: $\mathbf{e}_i = \frac{\mathbf{u}_i}{\|\mathbf{u}_i\|}$
4. The resulting set $\{\mathbf{e}_1, \mathbf{e}_2, \ldots, \mathbf{e}_n\}$ is an orthonormal basis for the span of the original set

Mathematical Properties and Theorems

• Gram-Schmidt process preserves the span of the original set of vectors
  • The orthonormal basis generated spans the same subspace as the input vectors
• The Gram-Schmidt process produces a unique orthonormal basis for a given input set
  • The order of the input vectors may affect the resulting basis, but the subspace spanned remains the same
• Orthogonal projection theorem: The projection of a vector onto a subspace is the closest point in the subspace to the vector
  • Minimizes the distance between the vector and its projection
• Pythagorean theorem in inner product spaces: For orthogonal vectors $\mathbf{u}$ and $\mathbf{v}$, $\|\mathbf{u} + \mathbf{v}\|^2 = \|\mathbf{u}\|^2 + \|\mathbf{v}\|^2$
• Parseval's identity: The sum of the squares of the inner products of a vector with an orthonormal basis equals the square of the vector's norm
  • $\sum_{i=1}^{n}|\langle \mathbf{v}, \mathbf{e}_i \rangle|^2 = \|\mathbf{v}\|^2$ for an orthonormal basis $\{\mathbf{e}_1, \mathbf{e}_2, \ldots, \mathbf{e}_n\}$

Practical Examples and Problem-Solving

• Example 1: Orthonormalizing a set of vectors in $\mathbb{R}^3$
  • Given vectors $\mathbf{v}_1 = (1, 0, 1)$, $\mathbf{v}_2 = (1, 1, 0)$, and $\mathbf{v}_3 = (1, 1, 1)$, find an orthonormal basis using the Gram-Schmidt process
• Example 2: Orthogonal decomposition of a vector
  • Given a vector $\mathbf{v} = (2, 3, 4)$ and an orthonormal basis $\{\mathbf{e}_1, \mathbf{e}_2, \mathbf{e}_3\}$, find the orthogonal decomposition of $\mathbf{v}$ with respect to the basis
• Example 3: Least squares approximation using orthogonal projections
  • Given a set of data points $(x_i, y_i)$ and a set of basis functions $\{f_1(x), f_2(x), \ldots, f_n(x)\}$, find the best linear combination of the basis functions that approximates the data using orthogonal projections
• Problem-solving strategy: Identify the input vectors, apply the Gram-Schmidt process step-by-step, and interpret the results in the context of the problem

Common Mistakes and Pitfalls

• Forgetting to normalize the orthogonal vectors at each step
  • Normalization ensures that the resulting vectors have unit length and form an orthonormal basis
• Incorrectly calculating the projections or subtractions
  • Double-check the formulas for projections and ensure that the subtractions are performed correctly
• Attempting to orthogonalize a linearly dependent set of vectors
  • The Gram-Schmidt process assumes that the input vectors are linearly independent
  • If the input vectors are linearly dependent, the process will encounter division by zero or produce non-orthogonal vectors
• Confusing orthogonality with linear independence
  • Orthogonality is a stronger condition than linear independence
  • An orthogonal set of vectors is always linearly independent, but a linearly independent set may not be orthogonal
• Misinterpreting the order of the resulting orthonormal basis
  • The order of the input vectors affects the order of the output basis, but not the subspace spanned

Advanced Topics and Extensions

• Gram-Schmidt process in infinite-dimensional Hilbert spaces
  • Extends the concept of orthonormalization to function spaces and sequences
• Modified Gram-Schmidt process: A numerically stable variant that updates the projections iteratively
  • Improves the accuracy and stability of the orthogonalization process in the presence of rounding errors
• Householder reflections: An alternative method for orthogonalization based on reflections across hyperplanes
  • Used in the QR decomposition and other numerical linear algebra algorithms
• Orthogonal polynomials: Polynomial sequences that are orthogonal with respect to a given inner product
  • Examples include Legendre polynomials, Chebyshev polynomials, and Hermite polynomials
• Orthogonal matrix factorizations: Decompositions of matrices into products of orthogonal matrices
  • Singular Value Decomposition (SVD), QR decomposition, and Schur decomposition

Review and Practice Problems

1. Prove that the Gram-Schmidt process produces an orthonormal basis for a given linearly independent input set.
2. Apply the Gram-Schmidt process to the vectors $\mathbf{v}_1 = (1, 1, 1)$, $\mathbf{v}_2 = (0, 1, 1)$, and $\mathbf{v}_3 = (0, 0, 1)$ to find an orthonormal basis for $\mathbb{R}^3$.
3. Given an orthonormal basis $\{\mathbf{e}_1, \mathbf{e}_2, \mathbf{e}_3\}$ and a vector $\mathbf{v} = (2, -1, 3)$, find the orthogonal projection of $\mathbf{v}$ onto the subspace spanned by $\mathbf{e}_1$ and $\mathbf{e}_2$.
4. Prove the Pythagorean theorem in inner product spaces for orthogonal vectors.
5. Explain why the Gram-Schmidt process may fail when applied to a linearly dependent set of vectors.
6. Compare and contrast the Gram-Schmidt process with the Householder reflections method for orthogonalization.
7. Discuss the importance of orthogonality in quantum mechanics and provide an example of an orthonormal basis in a Hilbert space.
8. Prove Parseval's identity for an orthonormal basis in an inner product space.