Linear Algebra in Physics
Editor's Note: This article bridges pure mathematics and physical theory. For computational implementations, see Matrix Methods in Computational Physics. For historical development, see History of Operator Theory.
Introduction
Linear algebra provides the fundamental mathematical language of modern physics. From the state vectors of quantum mechanics to the stress-energy tensors of general relativity, physical systems are routinely modeled as vector spaces equipped with linear operators, inner products, and metric structures. The discipline abstracts physical quantities into algebraic objects, enabling rigorous analysis, predictive modeling, and computational simulation across scales.
At its core, linear algebra studies vector spaces, linear transformations, and systems of linear equations. In physics, these structures map directly to observable quantities: positions and momenta become vectors in phase space, wavefunctions inhabit infinite-dimensional Hilbert spaces, and symmetries are represented by matrix groups.
Core Mathematical Structures
Vector Spaces & Bases
A physical state is typically represented as a vector \(\mathbf{v}\) in a vector space \(V\) over a field \(\mathbb{F}\) (usually \(\mathbb{R}\) or \(\mathbb{C}\)). The choice of basis \(\{\mathbf{e}_i\}\) allows decomposition into coordinates \(v^i\), where \(\mathbf{v} = \sum_i v^i \mathbf{e}_i\). In physics, basis transformations correspond to changes in reference frames or measurement contexts.
Linear Operators & Matrices
Physical observables and time evolution are encoded as linear operators \(\hat{A}: V \to V\). In finite dimensions, these correspond to matrices \(A^i_{\ j}\) satisfying \(\hat{A}\mathbf{v} = \sum_j A^i_{\ j} v^j \mathbf{e}_i\). Key properties include linearity \(\hat{A}(\alpha\mathbf{u} + \beta\mathbf{v}) = \alpha\hat{A}\mathbf{u} + \beta\hat{A}\mathbf{v}\), composition, and invertibility.
Eigenvalues & Eigenvectors
The spectral decomposition \(\hat{A}\mathbf{v}_\lambda = \lambda \mathbf{v}_\lambda\) identifies invariant directions under transformation. In quantum mechanics, eigenvalues correspond to measurable quantities, while eigenvectors represent stationary states. Hermitian operators guarantee real eigenvalues and orthogonal eigenvectors, ensuring physical observables yield real measurement outcomes.
Inner Product Spaces & Hilbert Spaces
Physical theories require a notion of angle and length, formalized via inner products \(\langle\mathbf{u}|\mathbf{v}\rangle\). Complete inner product spaces (Hilbert spaces) accommodate infinite-dimensional systems like wave mechanics, ensuring convergence of Fourier series and spectral expansions.
Applications in Physical Theory
Quantum Mechanics
Quantum theory is fundamentally linear. States are rays in a complex Hilbert space \(\mathcal{H}\), observables are self-adjoint operators, and time evolution follows the linear Schrödinger equation \(i\hbar \frac{d}{dt}|\psi(t)\rangle = \hat{H}|\psi(t)\rangle\). Matrix mechanics (Heisenberg) and wave mechanics (Schrödinger) are isomorphic representations of the same underlying linear structure.
- Bra-ket notation: Dirac's formalism abstracts vectors \(|\psi\rangle\) and duals \(\langle\phi|\) independent of representation.
- Tensor products: Composite systems live in \(\mathcal{H}_A \otimes \mathcal{H}_B\), enabling entanglement theory.
- Pauli & spin matrices: \(SU(2)\) representations model intrinsic angular momentum.
Classical Mechanics & Dynamical Systems
Phase space \(\Gamma \cong \mathbb{R}^{2n}\) carries a symplectic structure \(\omega\), a non-degenerate, skew-symmetric bilinear form. Linearized dynamics near equilibrium yield matrices governing stability (Lyapunov exponents). Hamilton's equations can be written compactly as \(\dot{\mathbf{z}} = J \nabla H(\mathbf{z})\), where \(J\) is the canonical symplectic matrix.
Electromagnetism & Continuum Physics
Maxwell's equations in vacuum combine into a single linear partial differential equation \(\Box A^\mu = 0\) (in Lorenz gauge). The electromagnetic field tensor \(F_{\mu\nu} = \partial_\mu A_\nu - \partial_\nu A_\mu\) is an antisymmetric rank-2 tensor, naturally manipulated using linear algebra and exterior calculus.
Relativity & Geometry
Special relativity utilizes Minkowski space \((\mathbb{R}^4, \eta_{\mu\nu})\) with metric signature \((-,+,+,+)\). Lorentz transformations preserve the inner product and form the group \(SO(1,3)\). General relativity generalizes this to curved manifolds, where linear algebra operates locally in tangent spaces via the metric \(g_{\mu\nu}\) and connection coefficients.
Numerical & Computational Aspects
Modern physics relies heavily on numerical linear algebra. Key algorithms include:
- Eigensolvers: Lanczos, Arnoldi, and Davidson methods for large sparse Hamiltonians.
- Linear systems: Conjugate gradient, GMRES, and sparse LU decompositions for PDE discretizations.
- Tensor networks: Matrix Product States (MPS) and Tensor Train decompositions for many-body quantum systems.
Numerical Stability Note: Ill-conditioned matrices (high condition number \(\kappa(A) = \|A\|\|A^{-1}\|\)) amplify rounding errors. Physicists routinely employ regularization, preconditioning, and arbitrary-precision arithmetic when solving inverse problems or near-degenerate systems.
References & Further Reading
- Dirac, P. A. M. (1930). The Principles of Quantum Mechanics. Oxford University Press.
- Goldstein, H., Poole, C., & Safko, J. (2002). Classical Mechanics (3rd ed.). Addison-Wesley.
- Tymoczko, J., & Biedenharn, L. C. (2016). "Matrix Mechanics and Linear Algebra Foundations". Reviews of Modern Physics, 88(3).
- Navarro, R. (2021). Linear Algebra for Physicists: From Vectors to Tensors. Cambridge University Press.