Direct Linear System Solvers
Solve linear systems \( Ax = b \) using direct elimination methods. These methods provide exact solutions (within machine precision) in finite steps.
Gaussian Elimination
Classical elimination method with forward elimination and back substitution
Complexity:O(n³)
Stability:Stable with pivoting
Key Features:
- Forward elimination
- Back substitution
- Partial pivoting
Process: Ax = b → Ux = c
Gauss Jordan
Modified Gaussian elimination that produces reduced row echelon form
Complexity:O(n³)
Stability:Stable
Key Features:
- Complete elimination
- Diagonal form
- Direct solution
Process: Ax = b → Ix = x
LU Decomposition
Factorizes matrix into lower and upper triangular matrices
Complexity:O(n³)
Stability:Very Stable
Key Features:
- Matrix factorization
- Multiple RHS
- Efficient for repeated solving
Process: A = LU, then Ly = b, Ux = y
About Direct Methods
Direct methods solve linear systems \( Ax = b \) by transforming the coefficient matrix through a finite sequence of operations. These methods are characterized by:
- Finite termination: Exact solution in finite steps
- Predictable cost: Known computational complexity
- Reliability: Work for most well-conditioned systems
- Memory efficiency: Can operate in-place