# Polynomial differentation as a linear transformation

• March 5th 2013, 06:48 AM
learning
Polynomial differentation as a linear transformation
Show that the only eigenvalue of the map D(f) = (d/dx)f is 0. Is D diagonalisable? and why?
• March 5th 2013, 12:50 PM
jakncoke
Re: Polynomial differentation as a linear transformation
I'm gonna assume this is a linear operator on a finite dimensional polynomial space.

so basically what this question is asking is, If you have a polynomial P(x) = $a_nx^n+...+a_0$ then $d\dx(P(x)) = n*a_nx^{n-1} +... + a_1$
when is applying this operator on a polynomial equal to multiplying the polynomial by a constant (eigen value)? Notice that d/dx reduces the degree of the polynomial if you have a non constant polynomial, so can you ever do $\lambda * P(x) = d/dx(P(x))$? If P(x) is non constant (meaning it has atleast 1 x term) then the left side would be of a higher degree than the right side. So this must possible only when p(x) is constant. when p(x) is constant then d/dx p(x) = 0. so $p(x) = a_0$ $d/dx(p(x)) = 0 = 0*p(x)$ so 0 is the only eigen value of the map. The eigen vectors are all the constant polynomials in the polynomial space.

A linear map is diagnizable iff the sum of dimensions of its eigen spaces is equal to dim V. So basically can we find n lin. independent eigen vectors for the linear map? If so then the linear map is diagnizable. Since we only found once eigen vector for the linear map, this matrix is not diagnizable. (Assume the polynomial space is not just a bunch of constants i.e has dim > 1).
• March 5th 2013, 11:50 PM
learning
Re: Polynomial differentation as a linear transformation