Differential Equations: The Language of Change

Ordinary Differential Equations

Differential equations: equations involving derivatives — describe how quantities change over time or space. ODE: one independent variable (usually time). y' = f(t, y). First order: population growth dy/dt = ry (exponential, Malthus), logistic growth dy/dt = ry(1−y/K) (Verhulst). Separable equations: dy/dx = f(x)g(y) → ∫dy/g(y) = ∫f(x)dx + C. Linear first-order: y' + P(x)y = Q(x) — solved by integrating factor μ = e^(∫P dx). Second order linear with constant coefficients: ay'' + by' + cy = 0 → characteristic equation ar² + br + c = 0. Harmonic oscillator: mx'' + bx' + kx = F(t) — models springs, circuits, bridges. Underdamped, critically damped, overdamped solutions. Existence and uniqueness (Picard-Lindelöf): if f is Lipschitz continuous, then IVP y'=f(t,y), y(t₀)=y₀ has a unique solution. Phase portraits: visualizing ODE systems in state space — equilibria, stability, limit cycles.

Partial Differential Equations

PDE: multiple independent variables — describe fields (temperature, pressure, electromagnetic). Heat equation: ∂u/∂t = α∇²u — diffusion of heat, solved by Fourier series (Fourier, 1822). Wave equation: ∂²u/∂t² = c²∇²u — vibrating strings, sound, electromagnetic waves. Laplace's equation: ∇²u = 0 — steady-state heat, electrostatics, potential theory. Schrödinger equation: iℏ∂ψ/∂t = Ĥψ — quantum mechanics foundation. Navier-Stokes equations: fluid dynamics — existence and smoothness of solutions is a $1M Millennium Prize Problem. Classification: elliptic (Laplace — steady-state), parabolic (heat — diffusion), hyperbolic (wave — propagation). Boundary conditions: Dirichlet (value specified), Neumann (derivative specified), Robin (linear combination). Separation of variables: assumes u(x,t) = X(x)T(t) — reduces PDE to ODEs. Green's functions: fundamental solutions — response to point source, enables superposition for any source distribution. Fourier transform: converts PDE in space to ODE in frequency domain — powerful for solving linear PDEs.

Chaos and Nonlinear Dynamics

Nonlinear systems: most real-world systems are nonlinear — superposition doesn't apply, qualitatively different behavior. Lorenz system (1963): three coupled ODEs modeling atmospheric convection — discovered sensitive dependence on initial conditions ("butterfly effect"). Strange attractors: fractal geometric structures in phase space — Lorenz attractor, Rössler attractor. Lyapunov exponents: quantify sensitivity to initial conditions — positive exponent = chaos. Bifurcation theory: qualitative changes in behavior as parameters vary — saddle-node, pitchfork, Hopf bifurcations. Logistic map: xₙ₊₁ = rxₙ(1−xₙ) — simple equation producing period doubling cascade to chaos. Feigenbaum constants: δ = 4.669..., universal ratios in period-doubling routes to chaos — appear in diverse systems. Fractal dimension: non-integer dimension of chaotic attractors — Mandelbrot set (boundary dimension ≈ 2), Koch snowflake (1.26). KAM theorem (Kolmogorov-Arnold-Moser): most tori in near-integrable Hamiltonian systems survive perturbation — stability in solar system. Three-body problem: generally chaotic — no general closed-form solution (Poincaré), requires numerical integration.

Numerical Methods and Applications

Euler's method: simplest numerical ODE solver — yₙ₊₁ = yₙ + hf(tₙ,yₙ). First-order accurate, unstable for stiff problems. Runge-Kutta methods: RK4 (classic 4th order) — the workhorse of numerical ODE solving. Adaptive step size: Dormand-Prince (ode45 in MATLAB) — adjusts h to control error. Stiff equations: implicit methods required (BDF, implicit Runge-Kutta) — chemical kinetics, electrical circuits. Finite difference methods (FDM): discretize PDE on a grid — replace derivatives with difference quotients. Finite element methods (FEM): divide domain into elements (triangles, tetrahedra), approximate solution with basis functions — COMSOL, ANSYS, FEniCS. Finite volume methods (FVM): conservation-based — standard in computational fluid dynamics (OpenFOAM, Fluent). Spectral methods: expand solution in basis functions (Fourier, Chebyshev) — exponential convergence for smooth problems. Physics-Informed Neural Networks (PINNs): neural networks trained to satisfy PDEs — encode physics as loss function. Applications: weather prediction (atmospheric PDEs), structural engineering (elasticity), electrodynamics (Maxwell's equations), financial mathematics (Black-Scholes PDE), epidemiology (SIR model), neuroscience (Hodgkin-Huxley equations).

❓ Frequently Asked Questions

Differential equations: equations involving derivatives — describe how quantities change over time or space. ODE: one independent variable (usually time). y' = f(t, y). First order: population growth ...

PDE: multiple independent variables — describe fields (temperature, pressure, electromagnetic). Heat equation: ∂u/∂t = α∇²u — diffusion of heat, solved by Fourier series (Fourier, 1822). Wave equation...

Nonlinear systems: most real-world systems are nonlinear — superposition doesn't apply, qualitatively different behavior. Lorenz system (1963): three coupled ODEs modeling atmospheric convection — dis...

Euler's method: simplest numerical ODE solver — yₙ₊₁ = yₙ + hf(tₙ,yₙ). First-order accurate, unstable for stiff problems. Runge-Kutta methods: RK4 (classic 4th order) — the workhorse of numerical ODE ...

Try it live

Everything above runs in your browser — open Differential Equations: The Language of Change and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.