Numerical Methods Summary

PDEs · Axiom Academy

SUMMARY Numerical Methods for PDEs Let's review the key approaches for solving partial differential equations computationally. Core Idea: Transform continuous PDEs into discrete algebraic equations that computers can solve Spatial Discretization: Divide the domain into a mesh or grid of discrete points Temporal Discretization: Break time into discrete steps for time-dependent problems Approximations: Replace derivatives with finite differences, integrals, or basis functions Trade-off: Finer discretization gives better accuracy but requires more computation Taylor Series Foundation: Approximate derivatives using neighboring grid points Forward Difference: (first-order accurate) Centered Difference: (second-order accurate) Laplacian Approximation: Second derivatives combine into the 5-point stencil Best For: Rectangular domains with regular grids Update Rule: Compute next time step directly from current values Advantages: Simple to implement, no matrix inversion required, computationally cheap per step Critical Limitation: Stability requires very small time steps (CFL condition) Efficiency: May need thousands of tiny steps for stability Use When: Spatial grid is coarse or when simplicity matters most Update Rule: Solve a system of equations at each time step Advantages: Unconditionally stable for many problems, allows large time steps Computational Cost: Requires matrix inversion or iterative solvers each step Efficiency: Fewer but more expensive time steps overall

This is the written version of the interactive lesson above. See the full PDEs course.