On a certain class of convex functions and the exact well-posedness classes of the Cauchy problem for the transport equation in Orlicz spaces (Q1288110)

The Cauchy problem is studied for a divergence-type first order linear equation \[ \rho _t + \text{div}(\rho\vec u) = h, \qquad\rho | _{t=0} = \rho_0, \] where \(t\in (0,T)\), \(0<T<\infty\), \(x=(x_1,\dots , x_n)\in \mathbb R^n\), \(n\geq 1.\) The class of these equations contains, for instance, the well-known continuity equation in continuum mechanics. Alongside this problem, the conjugate problem \[ \zeta_t + \vec u\cdot\nabla\zeta = g,\qquad\zeta | _{t=T} = \zeta_T \] is also studied. Here \(\rho\) and \(\zeta\) are unknown, whereas \(\vec u\), \(\zeta_T\), \(\rho_0\), \(g\), and \(h\) are given functions. For the sake of simplicity, all functions are assumed periodic in each space variable \(x_k\) with period \(T_k\). The main purpose of this article is to find minimal smoothness assumptions on \(\vec u = (u_1,\dots , u_n)\) which guarantee existence and uniqueness of solutions to the problem. The conditions on the quantity \(\text{div} \vec u\) are essentially important from the viewpoint of applications to the Navier-Stokes equation. In the article under review, the authors use Orlicz function spaces. This enables us to weaken the condition \(\text{div} \vec u\in L_1(0,T,L_{\infty}(\Omega))\) considered in the article by \textit{R. J. DiPerna} and \textit{P. L. Lions} [Invent. Math. 98, No. 3, 511-547 (1989; Zbl 0696.34049)] by replacing this condition with the integrability assumption for \(\text{div} \vec u\) in the corresponding spaces associated with convex functions \(M(s)\) in the class \(K\) of functions \(M\) such that \[ \int^{\infty}\frac{\ln M(s)}{s^2}ds = +\infty. \] Suitable counterexamples demonstrate that this condition is unimprovable for \(\text{div} \vec u\) to belong to the Orlicz space \(L_M(\Omega)\) with the functions \(M\) as above. The authors construct solutions in Orlicz spaces \(L_M\) and prove that these spaces are the exact well-posedness classes for the transport equations. The integral condition for the convex functions \(M(x)\) was used earlier by \textit{A. V. Kazhikhov} [Preprint / Tokyo Inst. of Tech., Ser. A, Math. Sci, A-119 (1996)] as a sufficient condition for constructing a solution to the problem and by \textit{A. E. Mamontov} [Preprint / Lavrentjev Inst. Hydrodynamics RAN, 2-96] for constructing global solutions to the equations for a compressible viscous fluid.