Linear Algebra
- Applications
- Matrices
- Numerical
- Decompositions
- Eigenvalues
- Special Matrices
- Determinants
- Schur Complements
- Resources
Applications
Linear Dynamical Systems
LQR Control
Kalman Filters
Probability
Forbenius Perrod theorem - there is a steady sta probability distribution
Quantum Mechanics
Least Squares Optimization
Boundary Value Problems
Filtering
Fourier transforms Wavelet decompositions PCA
Not sure how to arrange this hierarchy
Matrices
Kronecker products
Numerical
Decompositions
LU
SVD
https://peterbloem.nl/blog/pca-4
Jordan
QR
Eigenvector
Power method
Krylov
Conjugate Gradient
####
Eigenvalues
Resolvent
Characteristic polynomial
Special Matrices
import numpy as np
import scipy.linalg as linalg
from scipy.linalg import toeplitz
K = toeplitz([2,-1,0,0])
print(K)
print(linalg.inv(K))
from sympy import *
init_printing(use_unicode=True)
Ksym = Matrix(K)
print(Ksym.inv())
Determinants
Funny funny fellows indeed. Geometrically is the “volume” spanned by the columns. If the matrix represents a transformation, if is the factor of volume shrinkage of the transformation
A definition is an antisymetric recursive one. Why is this formula right? https://en.wikipedia.org/wiki/Laplace_expansion
det([a b; c d]) = ad-bc
Cramer’s rule gives a direct solution to the inverse of a matrix. https://en.wikipedia.org/wiki/Cramer%27s_rule Mainly useful in the 2x2 case
Facts:
- det(A) = prod of eigvals.
- det(AB) = det(A)det(B)
- det(A) = product of pivots in LU form. A more useful way of calulating than brute force
Charactersitic Polynomial = det(A - lam). The roots of this polynomial are eigenvalues.
Schur Complements
Domain Decomposition
H-Matrices
Infinite domains
Circuit equivalents
Resources
- Horn, Roger A.; Charles R. Johnson (1985). Matrix Analysis.
- COmputational Science and Engineering