On an example of a stochastic differential equation in infinite dimensions (Q4029015)

[For the entire collection see Zbl 0754.00020.] The paper discusses the solution of stochastic differential equations in infinite dimensions. In particular the equation \(dX=-\Lambda Xdt+\Sigma dW\), is considered when \(X\) takes values in one of the sequence spaces \(\ell^ p\), \(1\leq p\leq\infty\), or \(c_ 0\). In effect, one has a sequence of stochastic differential equations. The results are, however, fragmentary and an example based on the terms \(\{\lambda_ n\}\) and \(\{\sigma_ n\}\) of the sequences \(\Lambda\) and \(\Sigma\) indicates possible pathologies: if for all \(\varepsilon>0\) the series \(\sum\exp(- \varepsilon^ 2\lambda_ n/\sigma^ 2_ n)\) converges and \(\sup_ n\sigma_ n\left[{\log^ +\lambda_ n\over\lambda_ n}\right]^{1/2}<\infty\), then a solution of the above equation has a modification in \(c_ 0\) having locally bounded trajectories in this space. However, if also \(\limsup_{n\to\infty}\sigma_ n\left[{\log^ +\lambda_ n\over\lambda_ n}\right]^{1/2 }>0\), then such a solution does not solve the equation in a weak sense.