The relationship between two approaches to a generalized solution of the Hamilton--Jacobi equation (Q1874969)

This paper deals with the Cauchy problem for the Hamilton-Jacobi equation \[ \begin{aligned} \partial V(t, x)/\partial t+ H(t, x,\partial V(t, x)/\partial x)&= 0,\quad (t,x)\in G= (0, T)\times\mathbb{R}^n,\\ V(t, x) &= \varphi(x),\quad x\in\mathbb{R}^n.\end{aligned} \] The solution of this problem is used in optimal control theory and differential game theory. The problem is known to have no classical solutions in the general case, but there are several equivalent approaches to defining generalized solutions, in particular, viscous and minimax solutions. The aim of this paper is to clarify the relationship between the notions of minimax and generalized weak solutions for the case in which the Hamiltonian \(H(t,x,s)\) is concave and satisfies the Lipschitz condition with respect to the momenta. The main result is the generalized linearity of the solution operator of the Cauchy problem proved on the basis of the theory of minimax solutions. Moreover, the minimax solutions of the Cauchy problem are described by a formula of a special form. In the case of a nonautonomous Hamilton-Jacobi equation with Hamiltonian concave in \(s\), the author shows that continuous minimax and generalized weak solutions are the same. Moreover, he reduces the formula for solutions to an explicit form and shows that it is a generalization of the classical Lax-Oleinik formula.