Home Start Chapter 1

Symbols

List of Symbols and Notation

Rendered from uploaded LaTeX with KaTeX. No raw LaTeX displayed.

This list collects the notation used throughout the book. Some symbols are standard and may be reused locally when the context is clear. Chapter-specific notation is introduced again where it first appears. Boldface letters usually denote vectors or matrices, while calligraphic letters often denote operators, spaces, or residuals.

\begingroup \small \renewcommand{\arraystretch}{1.18} \setlength{\tabcolsep}{5pt}

Symbols table

SymbolMeaning
General mathematical notation
\N,\Z,\R,\CNatural numbers, integers, real numbers, complex numbers
iImaginary unit, i^2=-1, when used in complex analysis or Fourier formulas
x,y,zScalar variables, unless otherwise stated
\mathbf x,\mathbf y,\mathbf zVectors in finite-dimensional spaces
A,B,M,NMatrices or linear operators, depending on context
IIdentity matrix or identity operator
0Zero scalar, zero vector, or zero matrix, according to context
\delta_{ij}Kronecker delta, equal to 1 if i=j and 0 otherwise
\abs{x}Absolute value or modulus
\norm{x}Norm of a vector, matrix, function, or operator
\norm{x}_pVector p-norm
\norm{A}_pInduced matrix p-norm
\norm{f}_{L^p}L^p-norm of a function
\norm{u}_{H^m}Sobolev norm of order m
\ip{x}{y}Inner product
x^T yEuclidean inner product for real vectors
x^\ast yHermitian inner product for complex vectors
A^TTranspose of A
A^\astConjugate transpose of A
A^{-1}Inverse of a nonsingular matrix or operator
A^\daggerMoore--Penrose pseudoinverse
\rank(A)Rank of A
\range(A)Range or column space of A
\nullspace(A)Null space of A
\spann\{v_1,\ldots,v_k\}Linear span of the listed vectors
\diag(d_1,\ldots,d_n)Diagonal matrix with diagonal entries d_i
\tr(A), \trace(A)Trace of A
\det(A)Determinant of A
\lambda(A)Eigenvalue of A, when a generic eigenvalue is meant
\sigma(A)Spectrum of A, or singular values when context indicates SVD
\rho(A)Spectral radius, \rho(A)=\max_{\lambda\in\sigma(A)}|\lambda|
\kappa(A), \cond(A)Condition number of A
O(h^p)Quantity bounded by C h^p as h\to0
o(h^p)Quantity smaller than h^p asymptotically
\dd x, \dd tDifferential element in integration
Floating-point arithmetic and error analysis
\fl(x)Floating-point representation or rounded value of x
\betaBase of a floating-point system
pPrecision, usually the number of digits in the significand
eFloating-point exponent
uUnit roundoff
\epsmachMachine epsilon
\deltaRelative rounding error in the model \fl(x)=x(1+\delta)
\theta_nAccumulated floating-point perturbation after n operations
\gamma_nBound \gamma_n=nu/(1-nu) for accumulated roundoff
e_kError at iteration k, usually e_k=x_k-x_\ast
rResidual, often r=b-A\widetilde x or r=f-Au_h
\etaBackward error, residual indicator, or tolerance parameter
\tauTolerance, time-like variable, or fractional clock depending on context
\mathrm{tol}Numerical stopping tolerance
Nonlinear equations
f(x)=0Scalar nonlinear equation
F(x)=0Nonlinear system
x_\astExact root or solution
x_kApproximation at iteration k
J_F(x)Jacobian matrix of F at x
s_kNewton correction or search step
\alpha_kDamping or step-length parameter
\Phi(x)Merit function, often \Phi(x)=\frac12\norm{F(x)}^2
mMultiplicity of a root, or dimension depending on context
Interpolation and approximation
x_jInterpolation node
f_jFunction value f(x_j)
p_n(x)Interpolating polynomial of degree at most n
\ell_j(x)Lagrange cardinal polynomial
f[x_0,\ldots,x_k]Divided difference
w_jBarycentric interpolation weight or quadrature weight
T_k(x)Chebyshev polynomial of degree k
P_k(x)Legendre polynomial of degree k
B_{i,p}(x)B-spline basis function of degree p
\Pi_n fPolynomial projection or interpolation of f
\omega_n(x)Nodal polynomial \prod_{j=0}^{n}(x-x_j)
Linear systems and matrix factorizations
Ax=bLinear system
\widetilde xComputed approximation to x
r=b-A\widetilde xLinear-system residual
A=LULU factorization
P A=LULU factorization with row pivoting
A=QRQR factorization
A=LL^TCholesky factorization for symmetric positive definite A
LLower triangular matrix
UUpper triangular matrix, or solution variable in PDE discretization
QOrthogonal or unitary matrix
RUpper triangular matrix, residual, or objective residual depending on context
MPreconditioner or matrix splitting component
Iterative methods and Krylov spaces
A=M-NMatrix splitting
G=M^{-1}NStationary iteration matrix
D,L,UDiagonal, strictly lower, and strictly upper parts of a matrix
\omegaRelaxation parameter in SOR, or quadrature/convolution weight depending on context
\mathcal K_k(A,r_0)Krylov subspace \spann\{r_0,Ar_0,\ldots,A^{k-1}r_0\}
r_kResidual at iteration k
p_kSearch direction in CG or nonlinear conjugate gradient
\alpha_k,\beta_kCG or conjugate-gradient recurrence coefficients
V_kMatrix whose columns form an orthonormal Krylov basis
\overline H_kUpper Hessenberg matrix in Arnoldi/GMRES
PPreconditioner, projection, or polynomial depending on context
Eigenvalue problems
Ax=\lambda xEigenvalue problem
\lambdaEigenvalue
xEigenvector
\rho(x)Rayleigh quotient, \rho(x)=x^TAx/(x^Tx)
\sigmaShift in inverse iteration, or spectrum depending on context
Q_kR_kQR factorization at iteration k
A_kMatrix produced by QR iteration
HHessenberg matrix
TTridiagonal matrix, final time, or Chebyshev polynomial depending on context
Least squares and regularization
\min_x\norm{Ax-b}_2Linear least-squares problem
A^TAx=A^TbNormal equations
A=U\Sigma V^TSingular value decomposition
\sigma_iSingular value
\lambdaRegularization parameter or eigenvalue depending on context
A^\daggerPseudoinverse
\lambda\|x\|^2Tikhonov regularization term
r(x)Nonlinear least-squares residual vector
J(x)Jacobian of a nonlinear residual
Numerical differentiation
hStep size or mesh spacing
D_h fFinite-difference approximation to a derivative
\delta_xDifference operator in the x-direction
\delta_{xx}Second-difference operator
DDifferentiation matrix, derivative operator, or diagonal matrix depending on context
D^{(m)}m-th derivative matrix
x_j=\cos(j\pi/N)Chebyshev--Lobatto node
D_{ij}Entry of a differentiation matrix
Numerical integration and quadrature
I(f)Exact integral of f
Q(f)Quadrature approximation to I(f)
w_jQuadrature weight
\xi_jReference quadrature node
x_jPhysical quadrature node
T(h)Trapezoidal approximation with step h
S(h)Simpson approximation with step h
R_{k,j}Romberg extrapolation entry
EError estimate
B(p,q)Beta function
B_z(p,q)Incomplete beta function
Initial-value problems for ODEs
y'(t)=f(t,y(t))Ordinary differential equation in first-order form
t_nTime node
y_nNumerical approximation to y(t_n)
hTime step
k_iRunge--Kutta stage slope
A=(a_{ij}), b_i, c_iButcher tableau coefficients
R(z)Stability function of a one-step method
z=h\lambdaScaled eigenvalue for stability analysis
\lambdaTest-equation eigenvalue in y'=\lambda y
\Delta tTime step in time-dependent problems
\mathrm{LTE}Local truncation error
Boundary-value problems and Galerkin notation
[a,b]Spatial interval for one-dimensional BVPs
\alpha,\betaBoundary data or fractional orders depending on context
VFunction space, often H_0^1(a,b) or H_0^1(\Omega)
V_hFinite-dimensional approximation space
vTest function
u_h,y_hGalerkin or finite-element approximation
\phi_iBasis or test function
K_{ij}Stiffness matrix entry
M_{ij}Mass matrix entry
F_iLoad vector entry
a(u,v)Bilinear form
\ell(v)Linear functional
R(u;v)Weak residual
J(U)Jacobian of nonlinear algebraic system
Partial differential equations
\OmegaSpatial domain
\partial\OmegaBoundary of the domain
u(x,t)PDE solution
\nabla uGradient of u
\Delta uLaplacian of u
\kappaDiffusion coefficient
cWave speed
aAdvection velocity or interval endpoint depending on context
\mu=\kappa\Delta t/\Delta x^2Heat-equation mesh ratio
\nu=a\Delta t/\Delta xCourant number for advection
\lambda=c\Delta t/\Delta xCFL number for wave equation
U_i^nApproximation to u(x_i,t^n)
F_{j+1/2}Numerical flux at a cell interface
\overline U_j^nCell average in finite-volume method
\widehat fNumerical flux function or Fourier coefficient depending on context
Spectral and pseudospectral methods
\widehat u_kFourier coefficient of mode k
e^{ikx}Fourier basis function
kFourier wavenumber or iteration index depending on context
\operatorname{FFT}Fast Fourier transform
\operatorname{IFFT}Inverse fast Fourier transform
T_k(x)Chebyshev polynomial
P_k(x)Legendre polynomial
DSpectral differentiation matrix
D^2Spectral second-derivative matrix
\etaNormalized modal coordinate in filtering
\sigma(\eta)Spectral filter
KMaximum retained Fourier wavenumber, or stiffness matrix depending on context
NPolynomial degree, grid size, or number of time steps depending on context
Optimization and nonlinear systems
f(x)Objective function
\nabla f(x)Gradient of f
\nabla^2 f(x)Hessian of f
g_kGradient \nabla f(x_k), or constraint function depending on context
H_kApproximate inverse Hessian or Hessian-related matrix
B_kHessian approximation in quasi-Newton methods
s_k=x_{k+1}-x_kStep displacement
y_k=\nabla f(x_{k+1})-\nabla f(x_k)Gradient difference in quasi-Newton methods
\rho_kTrust-region ratio or reciprocal curvature 1/(y_k^Ts_k) depending on context
\Delta_kTrust-region radius
m_k(s)Local quadratic model in trust-region methods
c(x)=0Equality constraint
g(x)\le0Inequality constraint
\mathcal L(x,\lambda,\mu)Lagrangian
\lambdaEquality multiplier, regularization parameter, or eigenvalue depending on context
\muInequality multiplier, heat ratio, or memory parameter depending on context
\thetaParameter vector in inverse problems or residual minimization
J(\theta), \mathcal J(\theta)Parameter-dependent objective functional
J_Q(\theta), \mathcal J_Q(\theta)Quadrature approximation of an objective
Fractional calculus and fractional differential equations
\alphaFractional order, commonly 0<\alpha<1, or boundary datum depending on context
\rhoKatugampola deformation parameter or spectral radius depending on context
\Gamma(z)Gamma function
B(p,q)Beta function
B_z(p,q)Incomplete beta function
E_\alpha(z)One-parameter Mittag--Leffler function
E_{\alpha,\beta}(z)Two-parameter Mittag--Leffler function
I_{a+}^{\alpha}fLeft Riemann--Liouville fractional integral
I_{a+}^{\alpha,\rho}fKatugampola fractional integral
D_{a+}^{\alpha}fRiemann--Liouville fractional derivative
{}^{C}D_{a+}^{\alpha}fCaputo fractional derivative
{}^{GL}D_{a+}^{\alpha}fGrunwald--Letnikov fractional derivative
{}^{C}D_{a+}^{\alpha,\rho}fCaputo--Katugampola fractional derivative
{}^{CF}D_{a+}^{\alpha}fCaputo--Fabrizio derivative
{}^{ABC}D_{a+}^{\alpha}fAtangana--Baleanu Caputo-type derivative
\mathcal D_\rhoKatugampola differential operator t^{1-\rho}\frac{\dd}{\dd t}
(t-s)^{-\alpha}Caputo weakly singular memory kernel
(t^\rho-s^\rho)^{-\alpha}Caputo--Katugampola weakly singular memory kernel
t_nTime node in fractional time discretization
hUniform time step
y=t^\rhoStretched Katugampola time variable
h_yStep size in the stretched variable y=t^\rho
a_kL1 weight a_k=(k+1)^{1-\alpha}-k^{1-\alpha}
g_k^{(\alpha)}Grunwald binomial weight
\omega_kLubich convolution quadrature weight
\delta(\zeta)Linear multistep generating polynomial in Lubich convolution quadrature
\zetaGenerating-function variable
\tau_\alpha(t)Caputo fractional clock t^\alpha/\Gamma(1+\alpha)
\tau_{\alpha,\rho}(t)Caputo--Katugampola clock t^{\alpha\rho}/(\rho^\alpha\Gamma(1+\alpha))
\mathcal R[u]Fractional differential-equation residual
\theta_m^{(q)}Chebyshev--Lobatto panel node on [0,1]
s_{j,m}Physical panel node in a Chebyshev--Lobatto memory panel
\ell_m^{(q)}Lagrange cardinal polynomial on a q-degree panel
d_{m,r}^{(q)}Coefficient of \theta^r in (\ell_m^{(q)})'(\theta)
W_{j,m}^{C,q}(t)Caputo Chebyshev--Beta panel weight
W_{j,m}^{CK,q}(t)Caputo--Katugampola Chebyshev--Beta panel weight
I_r^C(a,b;t)Caputo beta moment on a panel
I_r^{CK}(a,b;t)Caputo--Katugampola beta moment on a panel
H_\ell^nFast-convolution history variable
w_\ell,\eta_\ellSum-of-exponentials weight and exponent
LShort-memory window length or lower triangular matrix depending on context
Frequently overloaded symbols
\alphaFractional order, line-search constant, boundary datum, or method coefficient depending on context
\betaFractional power exponent, CG coefficient, or method coefficient depending on context
\gammaNonlinear coefficient, accumulated roundoff bound, or method coefficient depending on context
\lambdaEigenvalue, test-equation parameter, regularization parameter, or Lagrange multiplier
\muHeat-equation mesh ratio, inequality multiplier, or measure-like parameter
\rhoSpectral radius, trust-region ratio, or Caputo--Katugampola deformation parameter
DDerivative matrix, diagonal matrix, differential operator, or domain-dependent operator
KStiffness matrix, maximum Fourier mode, or generic compact set
MMass matrix, preconditioner, or number of exponential terms in fast convolution
RResidual, upper triangular factor, stability function, or reaction term
TFinal time, Chebyshev polynomial symbol, or tridiagonal matrix
uUnit roundoff in floating-point analysis or solution variable in differential equations