Universal cusp scaling in random partitions (Q6189677)

The paper under review is to study the universal scaling limit of random partitions obeying the Schur measure which depends on two sets of generic parameters \(X=(x_i)_{i\in N}\) and \(Y = (y_i)_{i\in N}\). The universality of the eigenvalue statistics in the scaling limit is one of the key concepts in the study of random matrices. This paper extends one of authors' previous work for that the associated discrete kernel is asymptotic to the higher Airy kernel in the scaling limit for \(X= Y\) case. Proposition 1,1 (as Proposition 3.8) presents the asymptotic behavior of the Schur measure kernel as higher Airy kernel for even p and higher order analog of the Pearcey kernel for odd p; Proposition 1.2 (Proposition 3.10) gives the p-Airy kernel in terms of the k-th derivatives of the p-Airy functions as a Christoffel-Darboux-type formula; Proposition 1.3 (Proposition 5.10) gives the large gap asymptotics of the level spacing distribution which is consistent with the Forrester-Chen-Eriksen-Tracy [\textit{Y. Chen} et al., J. Phys. A, Math. Gen. 28, No. 7, L\, 207-L\, 211 (1995; Zbl 0827.15043)] conjecture. Section 2 starts with Schur measure \(\mu (\lambda)\) on the set of partitions in terms of Schur functions, and use the Miwa variables \[ t_n = \frac{1}{n} \sum_{i=1}^{\infty} x_i^n, \ \ \ \ \overline{t}_n = \frac{1}{n} \sum_{i=1}^{\infty} y_i^n. \] Correlation function \(\rho_k(W)\) and Wave functions \({\mathcal J}(z) = \sum_{n\in Z}J(n) z^n\) are reviewed, and the expression of the kernel using wave functions is given in Lemma 2.6 with a proof. Section 3 studies the differential and difference equations (3.1) for the wave functions in the scaling limit, and Lemma 3.1 presents the formally biorthonormal property of the scaled wave functions, Lemma 3.2 presents the simplified differential equations in the scaling limit for wave functions. Lemma 3.5 presents the formally biorthonormal property of the Airy functions, and the asymptotic behavior for different regimes. Proposition 3.8 and Proposition 3.10 are done via their previous paper with preparations in this section. Section 4 focuses on gap probability which is given by the Fredholm determinant, and the total derivative of the Fredholm determinant is given in terms of Hamiltonians in Lemma 4.3.where the Hamiltonians associated with the Fredholm determinants is given in Proposition 4.10 (Hamitonian system) and Hamiltonian equations for wave functions and parameter dependence and integral motions are given consecutively, where the partition function of the unitary matrix model can be expressed by the character expansion with another set of the Miwa variables with alternative sign \((-1)^{n-1}\). Section 5 studies level spacing distribution for odd p case, starting with the s-derivative of the auxiliary wave functions in Lemma 5.2 and the integral of motions for this special case. In subsection 5.1, Hamiltonian structure for this case \(H=H_1=H_2\) has the property that the s-derivative of the Hamiltonian \(H^{'} = -q p + 2R^2\) and the logarithmic derivative of the Fredholm determinant \(d \log F(s) = - 2H ds\). The strategy to derive the closed differential equations for the auxiliary functions \((u, v)\) is to rewrite all the bilinear forms \((q_k p_l)_{k, l =0, \cdots, p-1}\) in terms of \((u, v)\) recursively in Proposition 5.7 by the principle of mathematical induction, Proposition 5.8 shows that the auxiliary function \((u, v)\) obey the closed coupled nonlinear differential equations, and the number of differential equations agrees with the number of the auxiliary functions. Combining with previous results, the s-derivative of the Hamiltonian (\(\frac{d^2}{ds^2}\log F(s) =- 2H^{'}\)) behaves as \(O(s^{2/p})\) as \(s\to \infty\), hence the Fredholm determinant \(F(s)\) behaves as \(\exp \{-C_p s^{2/p +2}\}\). It seems no way to have a rigorous argument to determine the constant \(C_p\). For \(p=3\) case, even one can obtain an explicit form of differential equations, and it is not clear how to characterize the level spacing distribution. Authors made some interesting points to apply Riemann-Hilbert analysis as in [\textit{P. M. Bleher} and \textit{A. B. J. Kuijlaars}, Commun. Math. Phys. 270, No. 2, 481--517 (2007; Zbl 1126.82010)]. Appendix A analyzes the Hamiltonian system of a single time variable with explicit integral motion and nonlinear differential equations as well as asymptotic behavior and the boundary condition in Proposition A.1 and Proposition A.2. It would be interesting to know if there is a way to understand the constant \(C_p\) in this single time variable case for \(p=3\). The principle to understand the universal scaling limit, arguments and methods are mainly followed in authors [J. High Energy Phys. 2021, No. 7, Paper No. 100, 48 p. (2021; Zbl 1468.81085)] work. There are several typos in the paper, for instance in Proposition 4.10.