On the Cauchy problem for first order discontinuous ordinary differential equations (Q5956513)

The author studies the initial value problem NEWLINE\[NEWLINEx'(t)=f(t,x(t)) \text{ for a.e. }t\in [0,1], \quad x(0)=0,NEWLINE\]NEWLINE where \(f:[0,1] \times \mathbb{R} \to \mathbb{R}\) is not necessarily continuous. He gives conditions for the existence of extremal solutions (in the Carathéodory sense) \(x^\star\) and \(x^{\star\star}\) which satisfy \(x^\star \leq x\) and \(x^{\star\star} \geq x\) on \([0,1]\) for any other solution \(x\) to the problem. It is however striking how much do the results coincide with the details of the Filippov solutions (compare e.g. the conditions I--III introduced by the author and the conditions that guarantee the existence of Filippov solutions and their transversality at discontinuity manifolds). It is a pity that this paper (apparently also \textit{E. R. Hassan} and \textit{W. Rzymowski} [Nonlinear Anal., Theory Methods Appl. 37A, No. 8, 997-1017 (1999, Zbl 0949.34005)] and others) fails to recognize and highlight this link! It would be interesting to see to which extent the results of this body of theory can be in fact recovered with much less strain within the more general framework of Filippov solutions. In particular, the example proposed by the author corresponds to nonunique Filippov solutions that move away from a discontinuity manifold (in his case \(x=0\)) in both directions and therefore separate as the minimal and maximal solutions in the sense of the author.