Circular law, extreme singular values and potential theory (Q847420): Difference between revisions

Let \(\{X_{kj}\}, k,j = 1,\dots, n \) be a double array of independent and identically distributed complex random variables with mean zero \(EX_{11} = 0\) and variance \(E|X_{11}|^2 = 1\). Today, it is known that the distribution of complex eigenvalues of the matrix \(n^{-1/2}{\mathbf X} = n^{-1/2}(X_{kj})\) converges under some conditions to the so-called circular law, i. e. the uniform distribution over the unit disk in the complex plane. In this paper, it is proved under the conditions when \(EX_{11} = 0\), \(E|X_{11}|^2 = 1\) and the fourth moment \(E|X_{11}|^4 < \infty\) that the empirical spectral distribution function converges to the uniform distribution over the unit disk in two-dimensional space. The proof is based on \textit{V. L. Girko}'s idea [Theory Probab. Appl. 29, No.~4, 694--706 (1984; Zbl 0565.60034)] using the representation of the logarithmic potential of the two-dimensional empirical spectral distribution. The question, whether the finite fourth moment condition can be removed, remains still open.