A randomized algorithm for solving parameter-dependent linear matrix inequalities: computational experience

A randomized algorithm for solving parameter-dependent linear matrix inequalities is investigated on an example. Namely, the choice of a stopping rule and a tunable parameter is discussed. Although the algorithm does not have a stopping rule in its original form proposed by Polyak and his coworkers, a practical stopping rule can be implemented. The choice of a tunable parameter is important because it affects the running time of the algorithm.