A tropical analogue of the Pauli problem and a splitting of quasithermodynamics (Q382198)

The author first describes the notion of thermodynamic fluctuations starting from that of the thermodynamic system on a phase space \(\mathbb{R} ^{2(d+1)}(\beta ,E)\) where \(\beta \) (resp. \(E\)) is the set of intensive (resp. extensive) coordinates, through a reduction of a part of degree freedom of the system. Then he describes the tropical analogue of the Pauli problem. Taking a function \(\Phi (x)\) on \(\mathbb{R}^{n}\), he introduces its Legendre transform \(\widetilde{\Phi }(y)\) on \(\mathbb{R}^{n}\). The Pauli problem is as follows: Given real and smooth functions \(U(x)\) and \(W(y)\), construct a pair of complex functions \(\Phi (x)\), \(\widetilde{\Phi }(y)\) such that \(U(x)=\mathrm{Im} (\Phi (x))\), \(W(y)=\mathrm{Im}(\widetilde{\Phi }(y))\) and \[ \prod_{\alpha =1,\dots ,m}\frac{\partial }{\partial y_{i_{\alpha }}}\frac{\partial ^{2} \widetilde{\Phi }(y)}{\partial y_{j\partial y_{k}}}\mid _{y=0}=-D_{i_{m}}\dots D_{i_{2}}D_{i_{1}}([\Phi ^{\prime \prime }(x)]^{-1})_{j,k}\mid _{x=0}. \] The main result of the paper proves that in the 1D case the Pauli problem has exactly two canonical solutions, under some hypotheses on the structure of \(U(x)\) and \(W(y)\). The paper ends with some insights into the multi-dimensional case.