Heat equation with strongly inhomogeneous noise (Q2485768): Difference between revisions

The author considers the stochastic partial differential equation \[ \frac{\partial }{\partial t} u = \frac12 \frac{\partial^2 }{\partial x^2} u + b(t,x,u) \frac{\sigma(dt dx)}{dt dx}(t,x) + a(t,x,u) \frac{\partial^2 }{\partial t \partial x} w^\rho(t,x), \] \(t>0, x \in \mathbb R\), with initial condition \(u(0,x)=\eta(x)\). Here \(a,b: [0,\infty) \times \mathbb R \times \mathbb R \rightarrow \mathbb R\) and \(\eta: \mathbb R \rightarrow \mathbb R\) are continuous functions, \(\sigma\) and \(\rho\) are positive Radon measures on \([0,\infty) \times \mathbb R\), \(\sigma(dt dx)/dt dx\) is the Lebesgue density of \(\sigma(dt dx)\) and \(w^\rho: [0,\infty) \times \mathbb R \times \Omega \rightarrow \mathbb R\) is an inhomogeneous two-parameter Brownian motion on \([0,\infty) \times \mathbb R\) based on \(\rho\). In order to study the equation, it is understood as a stochastic integral equation using stochastic integrals involving martingale measures. The author gives an existence (strong and weak) and uniqueness result of continuous solutions under some assumptions on \(\sigma(dt,dx)\) and \(\rho(dt,dx)\). For instance, \(\rho(dt,dx)\) does not need to be absolutely continuous with respect to Lebesgue measure.