Uniqueness of generating Hamiltonians for topological Hamiltonian flows (Q360427)

A celebrated theorem of Gromov and Eliashberg states that if a uniform limit of symplectic diffeomorphisms is a diffeomorphism, then it is symplectic. This allows to consistently define symplectic homeomorphisms as homeomorphisms that are uniform limits of symplectic diffeomorphism. Then, a natural question to ask is whether there exists a good notion of Hamiltonian homeomorphism and continuous Hamiltonian isotopy. Such a definition has been recently proposed in [\textit{Y. G. Oh} and \textit{S. Müller}, J. Symplectic Geom. 5, No. 2, 167--219 (2007; Zbl 1144.37033)], as follows. A continuous isotopy \((h^t)_{t\in[0,1]}\) is called Hamitonian if there exists a family of functions \((H_t)_{t\in[0,1]}\) which is a limit for Hofer's norm of a sequence of smooth Hamiltonians whose flows uniformly converge to \(h^t\). The Hofer norm of a family of functions \((F_t)_{t\in[0,1]}\) is by definition \(\|F\|=\int_0^1(\max F_t-\min F_t)dt\). The Hamiltonian \(H\) is then called ``generator'' of the continuous Hamiltonian isotopy \(h^t\). In the present paper, the authors prove a foundational result for this theory: up to normalization, a continuous Hamiltonian isotopy admits a unique generator. This improves a result of \textit{C. Viterbo} where uniqueness was proved in the case of the uniform norm instead of the Hofer norm [Int. Math. Res. Not. 2006, No. 11, Article ID 34028, 9 p. (2006; Zbl 1108.53042)].