Ill-conditioned Matrix

Understanding Ill-Conditioned Matrices

An ill-conditioned matrix is a term used in numerical analysis to describe a matrix that is not suitable for certain numerical operations, particularly inversion and solving linear systems. The condition of a matrix is a measure of how sensitive the solution of a system of linear equations is to changes in the coefficients of the matrix or the right-hand side vector.

What is a Matrix Condition Number?

The condition number of a matrix is a numerical value that provides a measure of how well-conditioned or ill-conditioned a matrix is. It is defined as the ratio of the largest singular value of the matrix to the smallest singular value. For a matrix A, the condition number, often denoted as κ(A), is given by:

κ(A) = ||A|| * ||A-1||

where ||A|| is the norm of the matrix A and A-1 is the inverse of A. The norm used can vary, but the most common is the L2 norm, which corresponds to the largest and smallest singular values of the matrix.

Characteristics of Ill-Conditioned Matrices

Ill-conditioned matrices have a high condition number, which indicates that the range between the largest and smallest singular values is very large. This large disparity leads to numerical instability when performing calculations, such as solving systems of equations, because small changes in the input data or rounding errors can result in disproportionately large changes in the results.

For example, consider a system of linear equations Ax = b. If A is ill-conditioned, a slight perturbation in b can lead to a significantly different solution vector x. This sensitivity makes it difficult to trust the solutions obtained from computations involving ill-conditioned matrices.

Causes of Ill-Conditioning

Several factors can contribute to a matrix being ill-conditioned:

  • Scale disparity: When the columns (or rows) of a matrix have vastly different scales, the matrix may become ill-conditioned.
  • Near dependency: If the columns (or rows) of a matrix are nearly linearly dependent, the matrix is close to being singular, which can lead to ill-conditioning.
  • Poorly chosen basis: In some applications, the choice of basis for representing the problem can lead to an ill-conditioned matrix.

Dealing with Ill-Conditioned Matrices

There are several strategies to mitigate the issues caused by ill-conditioned matrices:

  • Rescaling: Adjusting the scales of the matrix's columns or rows can sometimes improve conditioning.
  • Regularization: Techniques such as Tikhonov regularization add a small value to the diagonal elements of the matrix, effectively improving its condition number.
  • Precision: Using higher precision arithmetic can reduce the impact of rounding errors and improve the stability of numerical computations.
  • Singular Value Decomposition (SVD): SVD can be used to analyze the matrix and perform computations in a way that minimizes the effects of ill-conditioning.


Ill-conditioned matrices pose a challenge in numerical computations due to their sensitivity to input data and rounding errors. Recognizing when a matrix is ill-conditioned is crucial for ensuring the accuracy and reliability of numerical solutions. By understanding the condition number and employing appropriate strategies, one can often overcome the difficulties associated with ill-conditioned matrices and obtain more stable and trustworthy results.

Please sign up or login with your details

Forgot password? Click here to reset