the FTCS scheme

Imagine watching heat spread through a metal rod minute by minute. You know the temperature everywhere right now; you want it one minute later. The most direct guess is: each spot's new temperature is its old temperature, nudged by how it compares to its two immediate neighbours — hotter spots cool toward the average, cooler spots warm up. FTCS is exactly this intuition turned into a formula for the heat (diffusion) equation u_t = alpha * u_xx.

The name spells out the recipe: Forward in Time, Central in Space. You approximate the time derivative u_t by a forward difference (u_j^{n+1} - u_j^n) / k — looking one step ahead — and the space derivative u_xx by the central second-difference (u_{j-1}^n - 2 u_j^n + u_{j+1}^n) / h^2 at the current time level n. Equating them and solving for the only unknown, the new value, gives u_j^{n+1} = u_j^n + r * (u_{j-1}^n - 2 u_j^n + u_{j+1}^n), where r = alpha * k / h^2. Because every term on the right is already known at level n, you simply sweep across the grid and read off the new row — no system to solve. This is what makes FTCS an EXPLICIT scheme: cheap, simple, one node at a time.

The catch is stability. FTCS is only CONDITIONALLY stable: von Neumann analysis shows it blows up unless r = alpha * k / h^2 is at most 1/2. So if you refine the space grid by halving h, you must cut the time step k by a factor of four to stay stable — a brutal penalty that makes FTCS expensive precisely when you want accuracy. Cross that limit and the solution does not just get inaccurate; it explodes into wild oscillations within a few steps. FTCS is the textbook first example of explicit time-stepping, prized for clarity and cursed by its stability restriction.