Separation Of Variables

Definition and Principle

Concept

Separation of variables: technique solving PDEs by expressing solution as product of single-variable functions. Converts PDE into ODEs by isolating variables. Reduces complexity: multidimensional problem → multiple one-dimensional problems.

Mathematical Formulation

Assume solution u(x,y,z,t) = X(x)Y(y)Z(z)T(t) for linear PDEs with separable variables. Substitute into PDE, divide by product, isolate terms: each term depends on a single variable, set equal to separation constant.

Underlying Principle

Independence of variables implies each term equals a constant. Leads to system of ODEs. Solutions combined linearly due to PDE linearity. Basis for eigenfunction expansions and superposition principle.

Applicability Conditions

Linearity

Method applies only to linear PDEs: no nonlinear terms or variable coefficients coupling variables non-separably.

Homogeneous Boundary Conditions

Best suited for homogeneous BCs: Dirichlet, Neumann, Robin types. Inhomogeneous BCs require transformation or superposition.

Geometry and Coordinates

Applicable in coordinate systems where PDE and BCs separate naturally: Cartesian, cylindrical, spherical. Geometry must allow separation of Laplacian or differential operators.

Variable Separability

PDE coefficients and domain must permit variables to be separated algebraically. Coupled variables or nonlinear interactions invalidate direct application.

Methodology

Step 1: Assume Product Solution

Postulate u(x,y,...) = X(x)Y(y)... for PDE u_t = F(u_x,u_y,...). Substitution reduces PDE to algebraic relation in separated variables.

Step 2: Substitute and Separate

Insert product form into PDE. Divide through by product u. Result: sum of single-variable functions equal zero. Each term set to constant.

Step 3: Derive ODEs

Obtain ordinary differential equations for each variable function with separation constants as parameters.

Step 4: Solve ODEs

Solve resulting ODEs with applicable boundary conditions. Solutions often trigonometric, exponential, Bessel, or Legendre functions depending on geometry.

Step 5: Superposition

Combine solutions linearly using eigenfunction expansions or Fourier series to satisfy initial conditions or nonhomogeneous terms.

Boundary Conditions

Types

Dirichlet: specified function values on boundary. Neumann: specified normal derivatives. Robin: linear combination of value and derivative.

Role in Eigenvalue Problems

BCs determine eigenvalues and eigenfunctions. Proper BCs ensure discrete spectra, orthogonality, completeness of basis functions.

Homogeneous vs. Inhomogeneous

Homogeneous BCs enable straightforward separation. Inhomogeneous BCs require modification via substitution or Green’s functions.

Physical Interpretation

BCs represent physical constraints: fixed temperature, insulated boundary, or convective heat loss in heat conduction problems.

Eigenvalue Problems

Definition

ODEs from separation lead to Sturm-Liouville problems. Eigenvalues: separation constants with nontrivial solutions. Eigenfunctions: corresponding variable functions.

Properties

Eigenvalues: countably infinite, real, positive or zero depending on problem. Eigenfunctions: orthogonal under weight function.

Orthogonality

Integral orthogonality enables expansion of arbitrary initial conditions in eigenfunction series.

Normalization

Eigenfunctions normalized to simplify coefficients in series expansions.

L[y] + λw(x)y = 0Boundary ConditionsEigenvalue λ_n, Eigenfunction y_n(x)Orthogonality: ∫ y_m(x) y_n(x) w(x) dx = 0 if m ≠ n 

Common Partial Differential Equations

Heat Equation

u_t = α² u_xx. Describes diffusion processes. Separation yields temporal exponential decay and spatial eigenfunctions.

Wave Equation

u_tt = c² u_xx. Models vibrations, waves. Separation produces harmonic temporal and spatial modes.

Laplace Equation

Δu = 0. Governs steady-state phenomena. Separation leads to harmonic functions depending on domain.

Helmholtz Equation

Δu + k² u = 0. Arises in frequency domain wave problems. Eigenvalue k² linked to resonance frequencies.

Fourier Series and Expansions

Role in Solutions

Eigenfunctions often trigonometric: sine and cosine functions forming Fourier series expansions of initial/boundary data.

Expansion Coefficients

Determined by orthogonality integrals projecting initial conditions onto eigenfunctions.

Convergence

Series converges to solution under conditions: piecewise continuity, bounded variation of initial data.

Generalized Fourier Series

Extension to non-trigonometric eigenfunctions: Bessel, Legendre series for cylindrical/spherical domains.

Examples

Heat Equation on a Rod

Domain: 0 < x < L, u_t = α² u_xx, BCs: u(0,t)=0, u(L,t)=0, IC: u(x,0)=f(x).

Assume u(x,t) = X(x)T(t)X'' + λX = 0, X(0)=0, X(L)=0T' + α² λ T = 0Eigenvalues: λ_n = (nπ/L)²Eigenfunctions: X_n(x) = sin(nπx/L)Solution: u(x,t) = Σ A_n e^{-α² λ_n t} sin(nπx/L) 

Wave Equation on a String

Domain: 0 < x < L, u_tt = c² u_xx, BCs: fixed ends, u(0,t)=0, u(L,t)=0, ICs: initial displacement and velocity.

Assume u(x,t) = X(x)T(t)X'' + λX = 0, X(0)=0, X(L)=0T'' + c² λ T = 0Eigenvalues: λ_n = (nπ/L)²Eigenfunctions: X_n(x) = sin(nπx/L)Solution: u(x,t) = Σ [A_n cos(c√λ_n t) + B_n sin(c√λ_n t)] sin(nπx/L) 

Laplace Equation in a Rectangle

u_xx + u_yy = 0, 0

Limitations and Extensions

Nonlinear PDEs

Method fails for nonlinear PDEs due to variable coupling. Requires alternative methods or linearization.

Inhomogeneous BCs and PDEs

Inhomogeneous terms complicate separation. Use superposition, Green’s functions, or transform methods.

Complex Geometries

Irregular domains not separable. Numerical or approximate methods preferred.

Extensions

Generalized separation: non-product solutions, transform methods, Lie symmetries extend applicability.

Numerical Approaches

Finite Difference and Element Methods

Numerical discretization of PDEs when analytical separation impossible. Approximate solutions on grids or meshes.

Comparison with Separation of Variables

Analytical method: exact, closed-form, limited scope. Numerical: approximate, general-purpose, computationally intensive.

Hybrid Methods

Use separation solutions as basis functions or benchmarks for numerical schemes.

Historical Context

Early Developments

Originated in 18th century with d’Alembert, Euler, Fourier. Fourier’s heat conduction work formalized method.

Evolution

Developed with Sturm-Liouville theory, eigenfunction expansions. Extended to spherical harmonics, special functions.

Modern Usage

Foundation for analytical PDE solutions, spectral methods, quantum mechanics, mathematical physics.

Significance

Key tool bridging PDE theory, functional analysis, applied mathematics, engineering.

"It is by separation of variables that the infinite complexity of partial differential equations becomes accessible to human understanding." -- Peter D. Lax

References

• Evans, L.C., Partial Differential Equations, Graduate Studies in Mathematics, Vol. 19, AMS, 1998, pp. 1–700.
• Strauss, W.A., Partial Differential Equations: An Introduction, Wiley, 2008, pp. 1–400.
• Haberman, R., Applied Partial Differential Equations, 5th Ed., Pearson, 2012, pp. 1–600.
• Carrier, G.F., Krook, M., Pearson, C.E., Functions of a Complex Variable: Theory and Technique, SIAM, 2005, pp. 1–550.
• Olver, P.J., Introduction to Partial Differential Equations, Springer, 2014, pp. 1–350.