Distance between the spectra of graphs with respect to normalized Laplacian spectra (Q2312707)

Distance between the spectra of two \(n\)-vertex graphs \(G\) and \(G^\prime\), with \(\lambda_1\geq\dots\geq\lambda_n\) and \(\lambda^\prime_1\geq\dots\geq\lambda^\prime_n\) as the eigenvalues of appropriate matrices, was defined by R. Brualdi in [\textit{D. Stevanović}, Linear Algebra Appl. 423, No. 1, 172--181 (2007; Zbl 1290.05002)] as \(\sum_{i=1}^n (\lambda_i-\lambda^\prime_i)^2\), while the cospectrality of \(G\) is defined as the smallest spectral distance to a graph not cospectral to \(G\). The authors here calculate spectral distance with respect to the normalized Laplacian matrix, determine cospectralities of \(nK_1\), \((n-2)K_1\cup K_2\) and \(K_n\), and show that cospectrality of any graph is at most four.