BACK TO ARCHIVESNumerical Methods
calendar_todayMar 2024folder_openComputing Principlesterminalnumerical analysis / scientific computing
Computational techniques for scientific computing.
Solving Linear Systems
Direct Methods
- Gaussian Elimination: Row reduction and backsubstitution
- LU Decomposition: Efficient for multiple right-hand sides
- Cholesky Decomposition: For symmetric positive-definite matrices
- QR Decomposition: Numerically stable
Iterative Methods
- Jacobi Iteration: Simple but slow
- Gauss-Seidel: Improved convergence
- SOR (Successive Over-Relaxation): Acceleration technique
- Conjugate Gradient: For symmetric positive-definite systems
Matrix Properties
- Condition number
- Matrix norms
- Ill-conditioned systems
Applications
- Finite element analysis
- Image processing
- Machine learning optimization
Numerical Stability
- Pivoting strategies
- Error propagation
- Algorithm selection