Condition Number + Upper Bounds

In summary, the conversation discusses the derivation of an upper bound on the infinity norm of a tridiagonal matrix, T_{n}, and the condition number of the matrix. The steps for finding the inverse of T_{n} and the upper bound for its infinity norm are shown. The condition number is also bounded by the infinity norm of the matrix and its inverse. Further clarification and explanation on the steps for finding these results is requested.
  • #1
twoski
181
2

Homework Statement



Given the tridiagonal matrix T[itex]_{n}[/itex]

2+x 1 0 0
1 2+x 1 0
0 1 2+x 1 etc.

Derive an upper bound on the infinity norm ||T[itex]_{n}^{-1}[/itex]|| and also derive an upper bound on the condition number of the matrix

The Attempt at a Solution



This is not actually homework, but rather a question from a study exam. I have the solution but the steps taken to get to the solution are really hard to follow so i was hoping to get some clarity.

First we note T[itex]_{n}[/itex] = (2+x)(I+B) where B contains all off-diagonal entries in [itex]_{n}[/itex]. This much makes sense to me.

Next we show that the inverse of [itex]_{n}[/itex] = 1 / (2+x) (I+B)[itex]^{-1}[/itex]. Generally speaking, do we put the off diagonal value from B in this equation? ie. If A = (9-x)(I+B) then i would say the inverse is 1/(9-x) (I+B)[itex]^{-1}[/itex], right?

The upper bound for the infinity norm of the inverse is ||T[itex]_{n}^{-1}[/itex]|| ≤ 1/(2+x) * 1/(1-(2/(2+x))) = 1/x, this doesn't really help me learn anything since these notes don't show how it's actually calculated. If anyone could show how i might go about doing this it would be appreciated.

The condition number is bounded by the infinity norm of the matrix multiplied by the infinity norm of its inverse which is less than or equal to (4+x)/x.
 
Physics news on Phys.org
  • #2
Again, i would really appreciate it if someone could show me the steps taken to get to these results. Thanks!
 

FAQ: Condition Number + Upper Bounds

What is a condition number?

The condition number is a measure of how sensitive a mathematical problem is to changes in the input data. It is commonly used in the context of numerical algorithms to determine the stability and accuracy of the solution.

How is the condition number calculated?

The condition number is typically calculated as the ratio between the largest and smallest possible changes in the output of a mathematical problem, divided by the largest and smallest possible changes in the input. This can be expressed as a single number or as a function of the input data.

Why is the condition number important?

The condition number is important because it indicates how well a mathematical problem can be solved numerically. A high condition number means that small changes in the input data can result in large changes in the output, making it difficult to obtain an accurate solution. A low condition number, on the other hand, means that the problem is well-posed and can be solved with greater precision.

What are upper bounds in relation to the condition number?

Upper bounds refer to the maximum possible value that a condition number can take. They are often used as a way to compare the condition number of different mathematical problems and determine which one is more well-posed or easier to solve numerically.

Can the condition number be used to predict the accuracy of a solution?

Yes, the condition number can be used to estimate the accuracy of a solution. A lower condition number typically means that the solution will be more accurate, while a higher condition number indicates that the solution may be less accurate or more sensitive to changes in the input data. However, the condition number is not the only factor that affects the accuracy of a solution, and other considerations such as the numerical method used and the quality of the input data should also be taken into account.

Back
Top