I was working on the following question and got stuck (Chap 2, #10 Kinderleher and Stampaccia)

Let $\Omega$ be a bounded domain with a smooth boundary $\partial \Omega$ and suppose given $f \in H^{-1}(\Omega)$ and $g \in L^2(\Omega)$. For $\lambda \in \mathbb{R}$ study the existence and uniqueness of solutions of the variational inequality:

$$u \in K : \int_\Omega[u_{x_i}(v-u)_{x_i} + \lambda(v-u)]\,dx \geq \langle f,v-u\rangle + \int_{\partial \Omega} g(v-u)\,dx \quad \forall v \in K$$

Where $K = \{v \in H^1(\Omega):v \geq 0\text{ on }\partial \Omega\}$. (Here I've used summation convention and partial derivatives are to be understood in the weak sense)

I think I have a lot of the pieces:

  1. The LHS, written as $a(u,v) = \int_\Omega[u_{x_i}(v)_{x_i} + \lambda(vu)]\,dx$ is coercive on $H^1(\Omega)$ when $\lambda >0$
  2. K is convex in $H^1$

So I think there will exist a weak solution to the Neumann Problem: $$-\Delta u + \lambda u = f \\ \frac{\partial u}{\partial n}=g$$ where $g \geq 0$.

I'm having trouble dealing with $f \in H^{-1}$ though. This $H^{-1}$ is dual to $H^1_0$, so the dual pairing on the RHS of the inequality only works when $v-u \in H^1_0$ But taking two elements arbitrary $v,u \in K$ doesn't necessarily give that $u-v$ is going to zero at boundary. This seems inconsistent and I'm having trouble resolving it.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.