Both Markovian and arbitrary residence time distributions are considered. The comparison of analytical predictions with the case of a time-decorrelated process shows that correlations can both decrease and increase the corresponding expected waiting time. Besides, the comparison of exponential, subexponential, and heavy-tailed models characterized by equal probabilities to observe the event of interest demonstrates that a faster decrease in the residence time probability density implies a shorter expected waiting time. Interestingly, irrespective of the details of a particular model for both discrete- and continuous-time jump processes considered here, the random waiting time becomes exponentially distributed in the long-time limit, thus, showing remarkable universality.

Optimization of the mean completion time of random processes by restart is a subject of active theoretical research in statistical physics and has long found practical application in computer science. Meanwhile, one of the key issues remains largely unsolved: how to construct a restart strategy for a process whose detailed statistics are unknown to ensure that the expected completion time will reduce? Addressing this query here we propose several constructive criteria for the effectiveness of various protocols of non-instantaneous restart in the mean completion time problem and in the success probability problem. Being expressed in terms of a small number of easily estimated statistical characteristics of the original process (MAD, median completion time, low-order statistical moments of completion time), these criteria allow informed restart decision based on partial information.