Themes
This page lists the main definitions, propositions, examples and skills that are relevant for the course. It is intended as a check list.
Do not forget the problem sets.
Sets and functions
Definitions
- of a set; cardinality (for finite and countable sets)
- unions, intersections, (relative) complements; cartesian products
- \(\mathbb N\), \(\mathbb Z\), \(\mathbb R\), \(\mathbb C\).
- of a function (domain, codomain, graph); range of a function
- surjectivity (onto), injectivity (one-to-one) and bijectivity (invertibility)
- the inverse of an invertible function
Skills/Methods
- to express a set using curly bracket-notation (\(\{ \}\)); to find the cardinality of a set; to determine intersections, unions and complements.
- to be able to use basic quantifiers.
- to understand functional notation; differences between domain/codomain, the function itself, and its values (\(f\colon X \to Y, x \mapsto f(x)\))
- to determine the range of a function.
- to determine whether a function, or some restriction of it, can be inverted (in reasonable cases: to find its inverse).
Examples
- A bijection \(\mathbb N \to \mathbb Z\); \(\mathbb N \to \mathbb N^2\).
- The operators \(1 \pm \partial_x^2\) on \(C_{2\pi-\text{per}}^\infty(\mathbb R, \mathbb R)\).
Metric spaces
Definitions
- of a metric; metric space
- induced metric; metric subspace
- \(B_r(x_0)\), \(S_r(x_0)\); the closed ball \(\tilde B_r(x_0)\)
- interior points, boundary points, closures; open and closed sets
- limits (continuous, sequential); accumulation points
- Cauchy sequences; completeness
- Isometries; embeddings
- Dense sets and separability
- Continuous mappings; contractions.
Theorems/propositions
- Normed spaces are metric spaces
- (Open) balls are open
- Uniqueness of limits; Sequential limits are as zero-limits for the distance function; Continuous and sequential limits agree
- Characterization of closures in terms of limits
- Convergent sequences are Cauchy; Cauchy sequences are bounded
- Characterization of complete subspaces of complete metric spaces
- The completion theorem
Skills/methods
- to understand the difference between a norm and a metric; between a normed space and a metric space.
- to determine whether a function defines a distance on some set \(X\).
- to determine whether a sequence is convergent; whether a sequence is Cauchy.
- to find the boundary of a set; to determine whether a set is open/closed.
- to know the main steps in the completeness proof for \(BC(I,\mathbb R)\).
Examples
- The discrete metric
- A metric space that is not a normed space
- To know that \(BC(I,\mathbb K)\) and \(l_p\) are complete.
- A non-trivial isometry
- To know that polynomials and \(C^k\)-functions are dense in \(BC\)-spaces (Stone–Weierstrass).
- A non-separable and some separable spaces.
- Completion of \(\mathbb Q\) in \(\mathbb R\); of \(C^\infty\)-functions in \(BC\)-spaces; of continuous functions in \(L_2\)-spaces.
Vector spaces
Definitions
- of a real (complex) vector space, subspace
- isomorphisms, embeddings
- linear combinations, spans, dependence, independence
- (Ordered) Hamel bases; dimension of a vector space
- Change-of-basis matrices
- Linear systems, (reduced) row echelon form, upper/lower triangular matrices, diagonal matrices; augmented matrix
- Linear transformations; the vector spaces \(L(X,Y)\) and \(L(X)\)
- Kernels and ranks; the null space, columns space and row space of a matrix
- Direct sums
- Nilpotent operators/matrices
- Convex sets
Propositions and theorems
- Linear span of a set is a subspace (smallest one containing the set).
- Finite-dimensional vector spaces are isomorphic to \(\mathbb R^n\) (\(\mathbb C^n\) if complex).
- Invertible matrices correspond to bases.
- \(L(X,Y)\) is a vector space.
- Linear transformations are determined by their action on any basis; finite-dimensional linear transformations correspond to matrices.
- The kernel and range of a linear transformation are vector spaces; a linear transformation is injective exactly if it has a trivial kernel.
- The rank–nullity theorem (algebraic and geometric version)
- Characterization of invertible linear transformations on a finite-dimensional vector space
- The Fredholm alternative
Skills/methods
- to understand the difference between a real and complex vector space.
- to know how to construct a vector space; determine whether a set is a subspace
- to determine linear dependence/independence
- to be able to switch from one basis to another (expressing vectors and linear transformations/matrices in the new basis).
- Gaussian elimination, LU-decompositions, Gauss-Jordan elimination.
- To calculate the determinant of a matrix.
Examples
- \(\mathbb R^n, \mathbb C^n\)
- \(C(I,\mathbb R), C(I,\mathbb C)\)
- \(P_n(\mathbb R)\), \(P(\mathbb R)\)
- \(P_n(\mathbb R) \cong \mathbb{R}^{n+1}\)
- \(P_n(\mathbb R) \hookrightarrow \mathbb{R}^{m}\) for \(m \geq n+1\); \(C^{m}(I,\mathbb R) \hookrightarrow C^n(I,\mathbb R)\) for \(m \geq n\).
- Linearly independent sets in \(\mathbb R^n\), \(\mathbb C^n\), \(l_2\) (real or complex), \(L_2((-\pi,\pi),\mathbb K)\) (real and complex), and \(P(\mathbb R)\).
- Canonical (standard) bases
- Dimensions of standard spaces
- The kernel and range of \(d/dx\) as an operator between different vector spaces.
Normed spaces
Definitions
- of a norm; normed space
- equivalence of norms
- of a Banach space
- isometrical isomorphisms
- (Ordered) Schauder bases
- Bounded linear transformations, the operator norm, and the space \(B(X,Y)\).
- Bounded linear functionals and the dual of a normed space.
Theorems and propositions
- Any two norms on \(\mathbb R^n\) are equivalent.
- Equivalent expressions for the operator norm.
- \(B(X,Y)\) is a normed space, Banach for \(Y\) Banach.
- A linear operator between normed spaces is continuous if and only if it is bounded.
- The kernel of a bounded linear operator is closed.
Skills/methods
- To determine whether a function is a norm.
- To use the axioms of a norm.
- To determine whether a sequence (vector) is in \(l_p\); a function is in \(BC(I,\mathbb R)\).
- To determine whether two norms are equivalent.
Examples
- \(l_p\)-spaces (including their norms; especially \(p=1,2, \infty\))
- \(BC(I,\mathbb R)\) (including the definition of the supremum norm \(\|\cdot\|_\infty)\)
- Standard Schauder bases for \(l_p\), and \(L_2\) on the interval \((-\pi,\pi)\)
- Integral operators as bounded linear transformations
- The dual of \(\mathbb R^n\); of \(L_2(I,\mathbb R)\).
- Finite-dimensional linear transformations are continuous.
Differential equations and spectral theory
Definitions
- Initial-value problems
- Lipschitz continuity (at least uniform Lipschitz continuity)
- Contractions
- Eigenvalues, eigenvectors and eigenspaces; spectrum and resolvent set.
- Characteristic polynomials; algebraic and geometric multiplicities; simplicity and semi-simplicity of eigenvalues.
- Fundamental systems/matrices.
- The matrix exponential; nilpotent matrices.
- Generalized eigenvectors and generalized eigenspaces; the maximal generalized eigenspace; the Riesz index.
- The transpose and conjugate transpose of a matrix; Hermitian (self-adjoint) matrices.
Theorems and propositions
- Higher-order ordinary differential equations can be reformulated as first-order systems.
- Lipschitz continuity is uniform on bounded and closed sets in \(\mathbb R^n\).
- The Banach fixed-point theorem (Picard iteration)
- The Picard–Lindelöf theorem
- The solution space of \(\dot x = Ax\) in \(\mathbb K^n\) is isomorphic to \(\mathbb K^n\).
- Properties of the matrix exponential
- The exponential solution formula for \(\dot x = Ax\), \(x(0) = x_0\).
- The Cayley–Hamilton theorem
- Each eigenvalue has a maximal generalized eigenspace.
Skills/methods
- To transform an ordinary differential equation into a first-order system
- To apply the Banach fixed-point theorem to solve fixed-point problems (such as the initial-value problem); to perform Picard iteration.
- To calculate the matrix exponential for a given matrix.
- To express a matrix in Jordan normal form.
- To solve the initial-value problem \(\dot x = A x\), \(x(t_0) =x_0\) in \(\mathbb R^n\).
Examples
- Examples showing that \(C^1(\mathbb R) \subsetneq Lip(\mathbb R) \subsetneq C^0(\mathbb R)\).
- On the spectral and Jordan decompositions.
Inner-product spaces
Definitions
- Inner products; inner-product spaces
- Hilbert spaces
- Orthogonality and orthogonal complements
- Orthogonal direct sums (\(\oplus\)) in Hilbert spaces
- For sequences and finite sets (systems): orthogonal, orthonormal, complete; Fourier coefficients and Fourier series
- Orthonormal bases
- Adjoints (transposes and conjugate transposes for matrices); self-adjoint operators (symmetric and Hermitian matrices).
- Unitary operators (unitary and orthogonal matrices)
- Positive definite and semi-definite matrices
Theorems and propositions
- The Riesz representation theorem
- Properties of an inner product (apart from the axiomatic properties)
- The Cauchy–Schwarz inequality
- Inner-product spaces are normed spaces.
- Relations between inner products and norms (the parallelogram law and polarization identity).
- The minimal distance theorem
- The projection theorem (corollary: strict subspace characterization)
- Fourier coefficients are best possible (corollary: closest point)
- The Fourier series theorem
- Properties of adjoints
- Self-adjoint operators have real spectrum and orthogonal eigenspaces.
- Unitary operators preserve inner products.
- Orthogonal matrices describe ortonormal bases.
- Characterization of positive definite matrices
Skills/methods
- To determine inner products, and whether a norm comes from an inner product
- To apply the Cauchy–Schwarz inequality
- To determine minimal distances and minimizing elements
- To understand and determine orthogonality in different inner-product spaces
- To apply the projection theorem and its corollary
- To determine/construct finite orthonormal sets.
- Gram–Schmidt orthogonalization and QR-decomposition
Examples
- Standard inner products on \(\mathbb R^n\) and \(\mathbb C^n\)
- Pythagoras theorem
- The Hilbert spaces \(l_2(\mathbb K)\) and \(L_2(I,\mathbb K)\).
- Pre-Hilbert spaces constructed using the \(L_2\)-norm (there are different ones, depending on which functions one starts with).
- Linear subspaces are convex sets.
- Minimization of distances in \(L_2((-\pi,\pi),\mathbb C)\) using the standard Fourier basis \(\{e^{ikx}\}_{k\in \mathbb Z}\)
- The Rank–nullity theorem in light of the projection theorem
- \(\overline{l_0} = l_2\) in \(l_2\).
- The duals of \(\mathbb R^n\), \(\mathbb C^n\) and \(L_2(I,\mathbb R)\)
- Orthonormal sequences and bases in standard Hilbert spaces
- Finding closest points using Fourier coefficients