A note on Wiener-Hopf determinants and the Borodin-Okounkov identity (Q1811168)

\textit{A. Borodin} and \textit{A. Okounkov} [Integral Equ. Oper. Theory 37, 386--396 (2000; Zbl 0970.47014)] proved an identity expressing a certain Toeplitz determinant as a Fredholm determinant. In~this note, the authors extend this to the Wiener-Hopf case. Namely, under appropriate hypotheses on a function \(F\in L^1(\mathbb{R})\) and on its Fourier transform \(K\), \[ \det(I-K_{[0,\alpha]}) = Z e^{c\alpha} \det(I-L_{[\alpha,\infty]}), \] where \(K_{[0,\alpha]}\) is the integral operator on \(L^2(0,\alpha)\) with kernel \(K(x-y)\), \(c\) and \(Z\) are certain constants, \(L_{[\alpha,\infty]}\) is the integral operator on \(L^2(\alpha,\infty)\) with kernel \[ L(x,y)=\int_0^\infty \Big(\frac{\phi_-}{\phi_+}-1\Big)_{x+z} \Big(\frac{\phi_+}{\phi_-}-1\Big)_{-y-z} \,dz, \] \((f)_x\) stands for the inverse Fourier transform of a function \(f\) at~\(x\), and \(1-F=\phi_+\phi_-\) is the Wiener-Hopf factorization of~\(1-F\). An application to an example arising in random matrix theory is also given.