Random Walk → Diffusion

The continuum limit: discrete lattice walks converge to the heat equation
⟨x²⟩ = 0 = Δx² · steps
Steps: 0 | D = Δx²/2Δt = 0.5
KS p-val vs Gaussian:

Central Limit Theorem → Heat Equation

A 1D random walk takes steps ±Δx each time Δt. After n steps, position X_n = Σ ξ_i where ξ_i = ±Δx with probability 1/2.

As Δx,Δt→0 with D = Δx²/(2Δt) fixed, the distribution P(x,t) converges to the diffusion equation: ∂P/∂t = D ∂²P/∂x² with solution P(x,t) = e^{-x²/4Dt}/√(4πDt).

This is the CLT in PDE form. The variance grows linearly: ⟨x²⟩ = 2Dt. Larger Δx → same continuum limit if D is preserved.

random walk diffusion equation CLT Brownian motion