How long does a diffusing particle take to reach a boundary? Exact theory vs. simulation.
For a 1D biased random walk on {0,…,N} with absorbing barriers, step right with probability p = ½+d, left with q = ½−d. Starting at x₀:
Drift-free (d=0): T(x₀) = x₀(N−x₀)
With drift: T(x₀) = (N/2d)·[(1−ρˣ⁰)/(1−ρᴺ)] − x₀/(2d), where ρ = q/p.
MFPT is a fundamental quantity in chemical kinetics, neural firing (integrate-and-fire), and financial ruin problems. The distribution is the inverse Gaussian.