On a nonlocal problem for a confined plasma in a tokamak. (Q2249855)

This article studies a class of nonlocal problems of the following form \[ \begin{cases} -\Delta u=f(x,u) \quad \text{in } \{u>0\},\\ -\Delta u=0 \quad \text{in } \{u\leq 0\},\\ u=\gamma \quad \text{(a constant to be determined) on } \partial \Omega ,\\ -\int_{\partial \Omega } \partial u / \partial \eta = I \text{ (a given constant)}, \end{cases}\tag{1} \] where \(\gamma <0\) and \(I>0\). This model explains the equilibrium of a plasma in a tokamak. Here, \(\Omega \) represents the cross section of the tokamak and it is a bounded, open and connected subset of \(\mathbb {R}^2\) with regular boundary \(\partial \Omega \). Special assumptions are made on the pressure term \(f(x,u)\), more precisely, \[ f(x,u)=\lambda \varphi (u_+)\Big | k( | \{u>u(x)\}| )\times u'_\ast (| \{u>u(x)\}| ) \Big | ^q+g(x), \] where \(1\leq q <2\), \(\lambda >0\) is the parameter which represents the ratio between the particle pressure and the magnetic pressure, \(\varphi :[0, +\infty ]\longrightarrow [0, +\infty ]\) is a continuous function, \(k\) is a function given by \(k(s)=\min \{s^{1/2}; (| \Omega | -s)^{1/2}\}\), the quantity \(| \{u>u(x)\}| \) denotes the measure of the set \(\{y\in \Omega : u(y)>u(x)\}\) and \(u_\ast \) is the decreasing rearrangement of \(u\) with \(u'_\ast (s)=du_\ast /ds\). Finally, \(g\geq 0\) is a positive function in \(L^r(\Omega)\) with \(r>1\). Under these assumptions, (1) can be rewritten in the following form: \[ \begin{cases} -\Delta u=\lambda \varphi (u_+)\Big | k( | \{u>u(x)\}| )\times u'_\ast (| \{u>u(x)\}| ) \Big | ^q 1_{u>0}+g(x)1_{u>0} \quad \text{in } \Omega ,\\ u=\gamma \quad \text{(a constant to be determined) on } \partial \Omega ,\\ -\int_{\partial \Omega } \partial u / \partial \eta = I \text{ (a given constant)}. \end{cases}\tag{2} \] The main theorem of this paper provides the existence of a solution \(u\in H^1(\Omega )\cap W^{2,\alpha }(\Omega)\) to (2) with \(\alpha =\min \{2/q, r\}\). Difficulties are related to the boundary conditions and to the nonlinear terms, and to overcome them, the authors introduce truncated problems in order to apply the Galerkin method.