-
Concordance probability in a big data setting: application in non-life insurance
Authors:
Robin Van Oirbeek,
Christopher Grumiau,
Tim Verdonck
Abstract:
The concordance probability or C-index is a popular measure to capture the discriminatory ability of a regression model. In this article, the definition of this measure is adapted to the specific needs of the frequency and severity model, typically used during the technical pricing of a non-life insurance product. Due to the typical large sample size of the frequency data in particular, two differ…
▽ More
The concordance probability or C-index is a popular measure to capture the discriminatory ability of a regression model. In this article, the definition of this measure is adapted to the specific needs of the frequency and severity model, typically used during the technical pricing of a non-life insurance product. Due to the typical large sample size of the frequency data in particular, two different adaptations of the estimation procedure of the concordance probability are presented. Note that the latter procedures can be applied to all different versions of the concordance probability.
△ Less
Submitted 14 November, 2019;
originally announced November 2019.
-
Multiple radial positive solutions of semilinear elliptic problems with Neumann boundary conditions
Authors:
Denis Bonheure,
Christopher Grumiau,
Christophe Troestler
Abstract:
Assuming $B_{R}$ is a ball in $\mathbb R^{N}$, we analyze the positive solutions of the problem \[
\begin{cases}
-Δu+u= |u|^{p-2}u, &\text{ in } B_{R},\newline
\partial_νu=0,&\text{ on } \partial B_{R},
\end{cases}
\] that branch out from the constant solution $u=1$ as $p$ grows from $2$ to $+\infty$. The non-zero constant positive solution is the unique positive solution for $p$ close t…
▽ More
Assuming $B_{R}$ is a ball in $\mathbb R^{N}$, we analyze the positive solutions of the problem \[
\begin{cases}
-Δu+u= |u|^{p-2}u, &\text{ in } B_{R},\newline
\partial_νu=0,&\text{ on } \partial B_{R},
\end{cases}
\] that branch out from the constant solution $u=1$ as $p$ grows from $2$ to $+\infty$. The non-zero constant positive solution is the unique positive solution for $p$ close to $2$. We show that there exist arbitrarily many positive solutions as $p\to\infty$ (in particular, for supercritical exponents) or as $R \to \infty$ for any fixed value of $p>2$, answering partially a conjecture in [Bonheure-Noris-Weth]. We give the explicit lower bounds for $p$ and $R$ so that a given number of solutions exist. The geometrical properties of those solutions are studied and illustrated numerically. Our simulations motivate additional conjectures. The structure of the least energy solutions (among all or only among radial solutions) and other related problems are also discussed.
△ Less
Submitted 17 March, 2016;
originally announced March 2016.
-
Asymptotic symmetries for fractional operators
Authors:
C. Grumiau,
M. Squassina,
C. Troestler
Abstract:
In this paper, we study equations driven by a non-local integrodifferential operator $\mathcal{L}_K$ with homogeneous Dirichlet boundary conditions. More precisely, we study the problem \[ \begin{aligned}
&- \mathcal{L}_K u + V(x)u = |u|^{p-2}u, &&\text{in } Ω, \newline
&u=0, &&\text{in } \mathbb{R}^N \setminus Ω, \end{aligned} \] where $2 < p < 2^{*}_s = \frac{2N}{N-2s}$, $Ω$ is an open bound…
▽ More
In this paper, we study equations driven by a non-local integrodifferential operator $\mathcal{L}_K$ with homogeneous Dirichlet boundary conditions. More precisely, we study the problem \[ \begin{aligned}
&- \mathcal{L}_K u + V(x)u = |u|^{p-2}u, &&\text{in } Ω, \newline
&u=0, &&\text{in } \mathbb{R}^N \setminus Ω, \end{aligned} \] where $2 < p < 2^{*}_s = \frac{2N}{N-2s}$, $Ω$ is an open bounded domain in $\mathbb{R}^{N}$ for $N\ge 2$ and $V$ is a $L^\infty$ potential such that $-\mathcal{L}_K + V$ is positive definite. As a particular case, we study the problem \[ \begin{aligned}
&(- Δ)^s u + V(x)u = |u|^{p-2}u, &&\text{in } Ω, \newline
&u=0, &&\text{in } \mathbb{R}^N \setminus Ω, \end{aligned} \] where $(-Δ)^s$ denotes the fractional Laplacian (with $0<s<1$). We give assumptions on $V$, $Ω$ and $K$ such that ground state solutions (resp. least energy nodal solutions) respect the symmetries of some first (resp. second) eigenfunctions of $-\mathcal{L}_K + V$, at least for $p$ close to $2$. We study the uniqueness, up to a multiplicative factor, of those types of solutions. The results extend those obtained for the local case.
△ Less
Submitted 16 July, 2014; v1 submitted 15 July, 2014;
originally announced July 2014.
-
Convergence of a mountain pass type algorithm for strongly indefinite problems and systems
Authors:
Christopher Grumiau,
Christophe Troestler
Abstract:
For a functional $\E$ and a peak selection that picks up a global maximum of $\E$ on varying cones, we study the convergence up to a subsequence to a critical point of the sequence generated by a mountain pass type algorithm. Moreover, by carefully choosing stepsizes, we establish the convergence of the whole sequence under a "localization" assumption on the critical point. We illustrate our resul…
▽ More
For a functional $\E$ and a peak selection that picks up a global maximum of $\E$ on varying cones, we study the convergence up to a subsequence to a critical point of the sequence generated by a mountain pass type algorithm. Moreover, by carefully choosing stepsizes, we establish the convergence of the whole sequence under a "localization" assumption on the critical point. We illustrate our results with two problems: an indefinite Schrödinger equation and a superlinear Schrödinger system.
△ Less
Submitted 8 January, 2013;
originally announced January 2013.
-
Nonlinear Schrödinger problems: symmetries of some variational solutions
Authors:
Christopher Grumiau
Abstract:
In this paper, we are interested in the nonlinear Schrödinger problem $-Δu + Vu = \abs{u}^{p-2}u$ submitted to the
Dirichlet boundary conditions. We consider $p>2$ and we are working with an open bounded domain $Ω\subset\IR^N$ ($N\geq 2$). Potential $V$ satisfies $\max(V,0)\in L^{N/2}(Ω)$ and $\min(V,0)\in L^{+\infty}(Ω)$. Moreover, $-Δ+ V$ is positive definite and has one and only one principal…
▽ More
In this paper, we are interested in the nonlinear Schrödinger problem $-Δu + Vu = \abs{u}^{p-2}u$ submitted to the
Dirichlet boundary conditions. We consider $p>2$ and we are working with an open bounded domain $Ω\subset\IR^N$ ($N\geq 2$). Potential $V$ satisfies $\max(V,0)\in L^{N/2}(Ω)$ and $\min(V,0)\in L^{+\infty}(Ω)$. Moreover, $-Δ+ V$ is positive definite and has one and only one principal eigenvalue. When $p\simeq 2$, we prove the uniqueness of the solution once we fix the projection on an eigenspace of $-Δ+ V$. It implies partial symmetries (or symmetry breaking) for ground state and least energy nodal solutions. In the litterature, the case $V\equiv 0$ has already been studied. Here, we generalize the technique at our case by pointing out and explaining differences. To finish, as illustration, we implement the (modified) mountain pass algorithm to work with $V$ negative, piecewise constant or not bounded. It permits us to exhibit direct examples where the solutions break down the symmetries of $V$.
△ Less
Submitted 20 December, 2012;
originally announced December 2012.
-
Lane Emden problems with large exponents and singular Liouville equations
Authors:
Massimo Grossi,
Christopher Grumiau,
Filomena Pacella
Abstract:
We consider the Lane-Emden Dirichlet problem -Δu = \abs{u}^{p-1}u, in B, u =0, on \partial B, where $p>1$ and $B$ denotes the unit ball in $\IR^2$. We study the asymptotic behavior of the least energy nodal radial solution $u_p$, as $p\rightarrow +\infty$. Assuming w.l.o.g. that $u_p(0) < 0$, we prove that a suitable rescaling of the negative part $u_p^-$ converges to the unique regular solution o…
▽ More
We consider the Lane-Emden Dirichlet problem -Δu = \abs{u}^{p-1}u, in B, u =0, on \partial B, where $p>1$ and $B$ denotes the unit ball in $\IR^2$. We study the asymptotic behavior of the least energy nodal radial solution $u_p$, as $p\rightarrow +\infty$. Assuming w.l.o.g. that $u_p(0) < 0$, we prove that a suitable rescaling of the negative part $u_p^-$ converges to the unique regular solution of the Liouville equation in $\IR^2$, while a suitable rescaling of the positive part $u_p^+$ converges to a (singular) solution of a singular Liouville equation in $\IR^2$. We also get exact asymptotic values for the $L^\infty$-norms of $u_p^-$ and $u_p^+$, as well as an asymptotic estimate of the energy. Finally, we have that the nodal line $\N_p:={x\in B : \abs{x}= r_p}$ shrinks to a point and we compute the rate of convergence of $r_p$.
△ Less
Submitted 7 February, 2013; v1 submitted 7 September, 2012;
originally announced September 2012.
-
Lane Emden problems: asymptotic behavior of low energy nodal solutions
Authors:
Massimo Grossi,
Christopher Grumiau,
Filomena Pacella
Abstract:
We study the nodal solutions of the Lane Emden Dirichlet problem $-Δu = |u|^{p-1}u with DBC on a smooth bounded domain $Ω$ in $\IR^2$ and where $p>1$. We consider solutions $u_p$ satisfying $p \int_Ω\abs{\nabla u_p}^2\to 16πe\quad\hbox{as}p\rightarrow+\infty\qquad (*)$ and we are interested in the shape and the asymptotic behavior as $p\rightarrow+\infty…
▽ More
We study the nodal solutions of the Lane Emden Dirichlet problem $-Δu = |u|^{p-1}u with DBC on a smooth bounded domain $Ω$ in $\IR^2$ and where $p>1$. We consider solutions $u_p$ satisfying $p \int_Ω\abs{\nabla u_p}^2\to 16πe\quad\hbox{as}p\rightarrow+\infty\qquad (*)$ and we are interested in the shape and the asymptotic behavior as $p\rightarrow+\infty$. First we prove that (*) holds for least energy nodal solutions. Then we obtain some estimates and the asymptotic profile of this kind of solutions. Finally, in some cases, we prove that $pu_p$ can be characterized as the difference of two Green's functions and the nodal line intersects the boundary of $Ω$, for large $p$.
△ Less
Submitted 18 June, 2012;
originally announced June 2012.
-
On the Mountain-pass algorithm for the quasi-linear Schrodinger equation
Authors:
Christopher Grumiau,
Marco Squassina,
Christophe Troestler
Abstract:
We discuss the application of the Mountain Pass algorithm to the so-called quasi-linear Schrodinger equation, which is naturally associated with a class of nonsmooth functionals so that the classical algorithm is not directly applicable.
We discuss the application of the Mountain Pass algorithm to the so-called quasi-linear Schrodinger equation, which is naturally associated with a class of nonsmooth functionals so that the classical algorithm is not directly applicable.
△ Less
Submitted 10 November, 2012; v1 submitted 2 May, 2012;
originally announced May 2012.
