Conditions for permanental processes to be unbounded (Q2406557)

The authors consider \(\alpha\)-permanental process \(X(t),\) which is a stochastic process with finite dimensional distributions that are \(\alpha\)-permanental vectors, i.e., having Laplace transform \(1/|I+KS|^{\alpha},\) where \(K\) and \(S\) are \(n\times n\) and diagonal matrices, respectively, \(\alpha>0.\) The permanent process is determined by a kernel \((K(s,t); s,t\in T)\) in a such way that \(K(t_i,t_j)\) determines the \(\alpha\)-permanental random variable by the latter Laplace transform. A concrete representation of permanental vectors, that is used to obtain a Sudakov type inequality that gives lower bounds for permanental processes that only requires that the inverse of the matrices \(K(t_i,t_j)\) are \(M\)-matrices, not symmetric, is given.