# Homework Help: Norm of matrix

1. Mar 2, 2008

### Gtay

1. The problem statement, all variables and given/known data
Let A = [a_{ij}] be a mxn matrix. Show that max$$_{ij}$$|a$$_{ij}$$| ≤ ‖A‖ ≤ √(∑$$_{ij}$$|a$$_{ij}$$)|

2. Relevant equations

3. The attempt at a solution

By the definition ‖A‖=max_{||x||≤1}‖A(x)‖ for all x ∈ Rⁿ.So, ‖A‖≥‖A∘(x₁,..,x_{n})$$^{T}$$‖ for x = (0,...,1,...0) with 1 is in the i$$^{ij}$$ position and so ‖A‖ ≥ ‖A∘(x₁,..,x_{n})$$^{T}$$‖ = ||(a$$_{i1}$$,a$$_{i2}$$,...,a$$_{ij}$$)|| = √(a$$_{i1}$$$$^{2}$$+...+a$$_{in}$$) ≥ max$$_{ij}$$|a$$_{ij}$$|.
I do not know what how to do the upper bound.

Last edited: Mar 2, 2008
2. Mar 2, 2008

Anyone?

3. Mar 2, 2008

### morphism

That's sort of hard to read - do you want to prove that $\| A \|^2 \leq \sum_{ij} |a_{ij}|^2$?

If so, the Cauchy-Schwarz inequality will be very useful.

4. Mar 2, 2008

### Gtay

Thank you for replying!
I will think about this.

Last edited: Mar 2, 2008
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook