Weak solutions to quasilinear wave equations of Klein-Gordon or sine-Gordon type and relaxation to reaction-diffusion equations (Q1376803): Difference between revisions

[Due to a technical error this article was first published with wrong mathematical signs. The correct article is distributed as a separatum] The author considers a quasilinear Klein-Gordon and corresponding sine-Gordon equation in one spatial dimension \[ \partial^2_tw-\partial_w \sigma (\partial_xw) +(1/ \tau) \partial_tw +f(w,x)=0; \quad w(0,x)= w_0(x),\;\partial_t w(0,x)= w_1(x). \] This could model the vibration of a non-uniform string with a damping term. He intends to prove existence of Lipschitz continuous global solutions. Introducing variables \(u=\partial_xw\), \(v=\partial_tw\), the author reduces the original equation to an equivalent system of equations. Shock waves in the system are assumed to travel at finite speed. They satisfy the Rankine-Hugoniot condition. This type of hyperbolic system without damping has been investigated by di Perna, who obtained bounds on the viscosity approximations to prove existence of global weak entropy solutions. The author states that these results could be directly obtained by using a version of Lax-Friedrichs or Godunov's fractional step finite difference schemes. He scales the variables \(u,v,w\) and considers convergence of approximate (finite difference) solutions, and as the relaxation time \(\tau \downarrow 0\), the scaled variable \(u^\tau\) converges strongly in \(L^2_{\text{loc}}\), \(\nu^\tau\) converges weakly in \(L^2_{\text{loc}}\), while \(w^\tau\) converges uniformly on compact sets to a function which satisfies the reaction-diffusion equation similar to the original one, but missing the damping term. Surprisingly the initial conditions have been lost in the limiting process, thus not affecting the final result. This 18 page article is well-written. All steps are explained, and for the benefit of readers who are not familiar with some developments the author defines carefully what is meant by weak solutions in the particular case under study, writes down the entropy inequality, gives a definition of weak entropy solution, and quotes the div-curl lemma of Tartar before using it. If a concept is used but is not defined or explained in this exposition, such as for example the Minty (also known as Dolph-Minty) monotonicity argument, the author supplies excellent references for the reader. In view of all this the reviewer forgives the author an occasional use of some rather strange typographic fonts.