Partial Differential Equations

A partial differential equation (PDE) is an equation relating an unknown function of several variables to its rates of change in those variables.

12 frameworks, from 1747 to today · Analysis · Mathematics

still a working view ended · each leads to its entry below

Overview

A partial differential equation (PDE) is an equation relating an unknown function of several variables to its rates of change in those variables1. Where an ordinary differential equation tracks a quantity changing along one parameter, a PDE describes a field—temperature across a metal plate, pressure in moving air, the shape of a stretched membrane, the electric potential around a charge1. Almost every quantitative science reaches a point where it must solve such equations, and the history of PDEs is the history of how mathematicians learned to pose them, solve them, and understand what their solutions can and cannot do1.

The subject's central difficulty is that a PDE rarely has a formula for its solution1. The work of the field is therefore threefold: proving that solutions exist and are unique, extracting their qualitative behavior (smoothness, oscillation, blow-up, propagation of disturbances), and computing them. Each era below is a shift in how one or more of these tasks was done.

Classical origins: characteristics, series, and the vibrating string

PDEs entered mathematics through physics in the eighteenth century. The wave equation for a vibrating string, the heat equation, and Laplace's equation for gravitational and electrostatic potential were written down by figures including d'Alembert, Euler, Daniel Bernoulli, and Lagrange. D'Alembert's 1747 treatment of the string introduced what became the method of characteristics: for a first-order equation, one can trace curves along which the equation reduces to an ordinary differential equation, so the solution is built by integrating along those curves. For the second-order wave equation the same idea yields d'Alembert's formula, expressing the solution as two traveling waves moving in opposite directions.

A famous dispute followed. Bernoulli argued the string's motion was a superposition of sine waves; Euler and d'Alembert objected that many physically reasonable initial shapes are not expressible as such series, and the argument forced mathematicians to ask what a function is and when a trigonometric series represents one. Fourier's work on heat conduction, published in 1822, settled the practical question by making Fourier series and Fourier integrals a general tool: decompose a field into simple oscillations, solve the equation separately for each, and reassemble. This spectral method remains fundamental, and its descendants—eigenfunction expansions, Fourier transform methods, separation of variables—are still the first technique taught for linear equations on simple domains.

Potential theory and Green's functions

Laplace's and Poisson's equations, governing gravitational and electrostatic potential, generated their own tradition23. Potential theory studies these solutions through integral representations: the potential at a point is a weighted average of sources, with a kernel that is singular at the source. George Green's 1828 essay introduced what we now call Green's functions—responses to a point source—and a formula converting a boundary-value problem into an integral over the boundary. The idea is that if you know how the domain responds to a unit impulse, you can build the response to any source by superposition4. This reframing turned hard boundary-value problems into integral equations and laid groundwork for both functional analysis and, much later, boundary element computation.

Rigor, distributions, and weak solutions

Nineteenth-century analysis largely treated PDE solutions as smooth functions given by formulas. The twentieth century replaced this with existence theory. The decisive move was to stop demanding a classical solution—one with continuous derivatives satisfying the equation pointwise—and instead ask for a weak solution: a function that satisfies the equation only after being tested against smooth functions and integrated by parts15. This requires the notion of a distribution (a generalized function), developed by Sobolev in the 1930s and given systematic form by Laurent Schwartz in the late 1940s67. Distributions let one differentiate functions that are not differentiable in the ordinary sense, and they made the Dirac delta and its derivatives legitimate objects7.

Weak solutions live naturally in Sobolev spaces, function spaces measuring size and derivatives in an averaged (integral) sense rather than pointwise8. Sobolev's embedding theorems, showing when such functions are actually continuous or bounded, became the working language of the field. The payoff was a general existence machinery: the Lax–Milgram theorem (1950s) guarantees solutions to a broad class of linear elliptic problems by reformulating them as questions about bounded operators on Hilbert spaces. The question of whether a weak solution is secretly smooth became the next great problem1.

Elliptic regularity

Elliptic equations—Laplace's equation and its generalizations—describe equilibrium states and are the best behaved19. The regularity program asked: if a weak solution exists, how smooth is it1? The answer, built through the mid-twentieth century by Schauder, Nirenberg, Agmon, Douglas, and others, is that elliptic equations smooth their solutions in the interior: a weak solution is as differentiable as the data allow, and singularities can occur only where coefficients or the boundary misbehave. The De Giorgi–Nash theorem (late 1950s), proved independently by Ennio De Giorgi and John Nash, extended this to nonlinear elliptic equations in divergence form, a result that stunned the field and became a model for later nonlinear regularity work. This theory is why elliptic problems are considered the tractable core of the subject9.

Parabolic theory and maximum principles

Parabolic equations, typified by the heat equation, describe diffusion and smoothing over time11011. Their characteristic tool is the maximum principle: a solution of the heat equation attains its extremes on the boundary of its space-time domain, never in the interior, which immediately gives uniqueness and stability. The principle was known in the nineteenth century and was generalized by Hopf and others in the twentieth. Parabolic regularity theory, developed alongside elliptic theory, shows that solutions become smooth for positive time even from rough initial data—diffusion erases detail1. These tools extend to nonlinear diffusion and to reaction-diffusion systems used across chemistry and biology.

Hyperbolic conservation laws and shocks

Hyperbolic equations describe wave propagation, and their history took a sharp turn with conservation laws—equations stating that a quantity such as mass or momentum is conserved. The difficulty is that smooth solutions can cease to exist in finite time: characteristics cross, and the solution develops a jump, a shock. Physically this is a sonic boom or a traffic jam; mathematically it means the equation must be interpreted in a weak sense, and weak solutions are not unique. The resolution, developed from the 1940s onward by Peter Lax, Olga Oleinik, James Glimm, and others, was to impose an entropy condition selecting the physically admissible shock. Glimm's 1965 existence proof for systems of conservation laws and the later development of entropy solutions and compensated compactness made rigorous sense of shock dynamics. This remains one of the hardest and most active areas, with direct application to gas dynamics and numerical simulation.

Variational methods and the direct method

Many PDEs are Euler–Lagrange equations of an energy functional: solving the equation is equivalent to minimizing energy12. The direct method in the calculus of variations, associated with Hilbert and refined by Tonelli and Morrey, seeks a minimizer by taking a minimizing sequence and proving it converges to something that is itself a minimizer. The obstacle is that minimizers may fail to be smooth, or may not exist among classical functions. This motivated geometric measure theory, developed by De Giorgi, Federer, and Fleming from the 1950s, which extends surfaces to currents and rectifiable sets so that minimal surfaces exist even with singularities. The payoff includes a rigorous theory of soap films and Plateau's problem—finding the surface of least area spanning a given boundary—and a general framework for variational problems with rough solutions1314.

Numerical computation: finite differences and finite elements

Because closed formulas are rare, computation became a branch of the subject in its own right1. Finite difference methods, dating to the early twentieth century and systematized by Courant, Friedrichs, and Lewy in 1928, replace derivatives by differences on a grid. Their central theorem—the CFL condition—states that a scheme converges only if information propagates no faster than the grid allows, a stability requirement that remains the first thing checked in any simulation.

Finite element methods, developed from the 1940s through the 1960s by engineers including Courant, Clough, and Zienkiewicz and given mathematical foundation by Babuška and others, instead approximate the solution by simple piecewise functions on a mesh and enforce the weak formulation. Their great advantage is handling complicated geometries and irregular boundaries, and their error theory rests directly on Sobolev spaces. Finite elements are now the dominant method for elliptic and parabolic problems in engineering, while finite differences and finite volumes dominate hyperbolic computation.

Nonlinear PDE and singularity analysis

The late twentieth century turned to genuinely nonlinear equations, where solutions can blow up, concentrate, or form singularities. The Navier–Stokes equations of fluid flow pose the most famous open problem: whether smooth solutions in three dimensions can develop singularities in finite time, one of the Clay Millennium Prize problems. Related work on nonlinear wave equations, the Korteweg–de Vries equation and solitons, and nonlinear Schrödinger equations revealed rich structure—integrable systems solvable by inverse scattering, and concentration phenomena where solutions focus on lower-dimensional sets. Singularity analysis, whether by energy methods, blow-up rescaling, or monotonicity formulas, became a discipline of its own, connecting PDEs to geometry through equations such as the Ricci flow and the minimal surface equation.

The present

Today the field is organized less by equation type than by method and by the tension between existence, regularity, and computation. Weak solutions and Sobolev spaces remain the universal language; elliptic and parabolic regularity theory the reliable core; hyperbolic conservation laws and nonlinear singularity formation the frontier. Numerical analysis is fully integrated, with rigorous a posteriori error control and the analysis of structure-preserving schemes. Open questions cluster around the Navier–Stokes regularity problem, the behavior of solutions to nonlinear systems in several dimensions, and the rigorous justification of continuum models derived from microscopic physics. The historical arc—from formulas for special equations, through existence theory for weak solutions, to the qualitative and computational study of nonlinear fields—has left the subject with a mature toolkit and problems that still resist it.

References

  1. Partial differential equation (introduction, Definition, Introduction and examples, Weak solutions, Regularity, Second order equations), Wikipedia
  2. Potential theory, Wikipedia
  3. Laplace's equation, Wikipedia
  4. Green's function, Wikipedia
  5. Weak solution, Wikipedia
  6. Sergei Sobolev, Wikipedia
  7. Distribution (mathematical analysis), Wikipedia
  8. Sobolev space, Wikipedia
  9. Elliptic operator, Wikipedia
  10. Parabolic partial differential equation, Wikipedia
  11. Heat equation, Wikipedia
  12. Calculus of variations, Wikipedia
  13. Plateau's problem, Wikipedia
  14. Geometric measure theory (History), Wikipedia

The frameworks of Partial Differential Equations

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1747-1900

    Method of Characteristics

    How do disturbances propagate along curves?

    Fly to its tower in Nousfarer

  2. 1822-Present

    Spectral method

    How can a field be built from simple oscillations?

    Fly to its tower in Nousfarer

  3. 1828-Present

    Potential theory

    How does a source determine a potential everywhere?

    Fly to its tower in Nousfarer

  4. 1900-Present
  5. 1900-Present

    Maximum Principles and Parabolic Theory

    Where can a diffusing quantity reach its extremes?

    Fly to its tower in Nousfarer

  6. 1928-Present

    Finite difference method

    When does a grid calculation converge to the true solution?

    Fly to its tower in Nousfarer

  7. 1930-Present

    Elliptic Regularity Theory

    How smooth must a solution of an equilibrium equation be?

    Fly to its tower in Nousfarer

  8. 1930-Present

    Weak solution

    What does it mean to solve an equation without differentiability?

    Fly to its tower in Nousfarer

  9. 1940-Present

    Hyperbolic Conservation Laws and Shock Theory

    How do discontinuities form and evolve in wave motion?

    Fly to its tower in Nousfarer

  10. 1945-Present

    Theory of distributions

    How can we differentiate functions that are not differentiable?

    Fly to its tower in Nousfarer

  11. 1950-Present

    Finite element method

    How can complex domains be approximated by simple pieces?

    Fly to its tower in Nousfarer

  12. 1960-Present

    Nonlinear Singularity and Blow-up Analysis

    When and how do nonlinear equations develop singularities?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Method of CharacteristicsSpectral methodPotential theoryCalculus of Variations and Geometric …Maximum Principles and Parabolic Theo…Finite difference methodElliptic Regularity TheoryWeak solutionHyperbolic Conservation Laws and Shoc…Theory of distributionsFinite element methodNonlinear Singularity and Blow-up Ana…

Beyond the field

Where Partial Differential Equations's ideas came from in other fields, and what they led to there: the strongest 12 of 65 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Partial Differential Equations is a planet in Nousfarer

Its city has a tower for each of these 12 frameworks. Land, walk in, and climb one concept at a time.

Land on Partial Differential Equations

Related fields

Fields its ideas are linked to

More in Analysis

See where Partial Differential Equations sits on the Map of Human Knowledge