Let $P(x)$ be a polynomial that satisfies: $P(0) = f(0)$, $P(2) = f(2)$, $P^{'} (2) = f^{'}(2)$
(a) Using Hermite interpolation, find $P(x)$.
(b) Compute the expression $\int^∞_0 e^{−x}P(x)dx$ using the approximation from part (a).
Proof:
Using the Hermite formula, it seems that the minimum degree we can compute using the 3 given values for 2 points is $2$. However, the formula gives a 3rd degree polynomial and not a 2nd degree polynomial. I need the other missing derivative value. How do I get around that?
Concretely, Since the Hermite polynomial is $H_n(x)=\sum_{k=1}^n f(x_k)h_k(x) +\sum_{k=1}^n f^{'}(x_k)\hat{h}_k(x)$ where
$$h_k=[1-2L_k^{'}(x_k)(x-x_k)](L_k(x))^2$$ and $$\hat{h}_k=[1-2L_k^{'}(x_k)(x-x_k)](L_k(x))^2$$
where
$$L_k(x)=\prod_{j=1,j\neq k}^n \frac{x-x_j}{x_k-x_j}$$
It is easy to see for instance that $L_1(x)=-\frac{x-2}{2}$ and $L^{'}_1(0)=-1/2$ thus $h_1(x)=\frac{1}{4}(1+x)(x-2)^2$ Which is already a 3rd degree polynomial. Not a 2nd degree. Which means I would need to know also $P^{'}(0)$ which I do not have. Can I use the Hermite Interpolating Theorem to build a second degree polynomial? What am I missing here?
I read here about specialized Hermite interpolation but there are no details.
Thanks and regards,