Abstract:
Poisson restart assumes that a stochastic process is interrupted and starts again at random time moments. A number of studies have demonstrated that this strategy may minimize the expected completion time in some classes of random search tasks. What is more, it turned out that under optimally tuned restart rate, any stochastic process, regardless of its nature and statistical details, satisfies a number of universal relations for the statistical moments of completion time. In this paper, we describe several new universal properties of optimally restarted processes. Also we obtain a universal inequality for the quadratic statistical moments of completion time in the optimization problem where stochastic process has several possible completion scenarios.
Fulltext:
PDF file (169 kB)
First page:
PDF file