. and * T. Rotbart and S. Reuveni had equal contribution to this work.We study the effect of restart, and retry, on the mean completion time of a generic process. The need to do so arises in various branches of the sciences and we show that it can naturally be addressed by taking advantage of the classical reaction scheme of Michaelis & Menten. Stopping a process in its midst-only to start it all over again-may prolong, leave unchanged, or even shorten the time taken for its completion. Here we are interested in the optimal restart problem, i.e., in finding a restart rate which brings the mean completion time of a process to a minimum. We derive the governing equation for this problem and show that it is exactly solvable in cases of particular interest. We then continue to discover regimes at which solutions to the problem take on universal, details independent, forms which further give rise to optimal scaling laws. The formalism we develop, and the results obtained, can be utilized when optimizing stochastic search processes and randomized computer algorithms. An immediate connection with kinetic proofreading is also noted and discussed.When engaged in a specific task for a time period that extends beyond our initial expectations, we are constantly faced with two alternatives-either keep on going or stop everything and start anew. Every now and then we opt for the latter, hoping that a fresh start will break-off an unproductive course of action and expedite the completion of the task at hand. This decision could, however, turn out to be counter-productive-nipping an awaited, but unforeseen, finale in the bud. To restart, or not to restart, that is therefore the question.Not at all unique to our everyday lives, a "dilemma" similar to the one described above is relevant to virtually any physical, chemical, or biological process that can be restarted. Most notably, restart (or unbinding) is an integral part of the renown Michaelis-Menten Reaction Scheme (MMRS) illustrated in Fig. 1 [1]. Originally devised to describe enzymatic catalysis, the MMRS has attracted on-growing scientific interest for more than a century [2]. Indeed, nature is full with an astonishing variety of Michaelian processes. DNA-DNA hybridization, antigen-antibody binding, and various other molecular processes can all be described by the MMRS [3]. That and more, the simplicity and generality of the scheme have rendered it widely applicable and it is now used to describe anything from heterogeneous catalysis [4-6] to in vivo target search kinetics [7]. As a matter of fact, one can easily convince himself that any first passage time (FPT) process [8]-be it the time to target of a simple Brownian particle, or that of more sophisticated stochastic processes [9][10][11][12] and random searchers [13][14][15]-can become subject to restart [16][17][18][19][20][21][22] and is then naturally accommodated by the MMRS. Wishing to acquire a unified view on restart phenomena we identify the MMRS as an ideal object of study.Central to our understanding of the ...