We consider a single-server queue with renewal arrivals and i.i.d. service times, in which the server employs either the preemptive Shortest Remaining Processing Time (SRPT) policy, or its non-preemptive variant, Shortest Job First (SJF). We show that for given stochastic primitives (initial condition, arrival and service processes), the model has the same fluid limit under either policy. In particular, we conclude that the well-known queue length optimality of preemptive SRPT is also achieved, asymptotically on fluid scale, by the simpler-to-implement SJF policy. We also conclude that on fluid scale, SJF and SRPT achieve the same performance with respect to response times of the longest-waiting jobs in the system.
Background: The U.S. Food and Drug Administration (FDA) conducted a benefit–risk assessment for Moderna’s COVID vaccine mRNA-1273 prior to its full approval, announced 31 January 2022. The FDA’s assessment focused on males 18–64 years old because its risk analysis was limited to vaccine-attributable myocarditis/pericarditis (VAM/P), given the excess risk among males. The FDA’s analysis concluded that vaccine benefits outweighed risks, even for 18–25-year-old males (those at highest VAM/P risk). We reanalyze the FDA’s benefit–risk assessment using information available through the third week of January 2022 and focusing on 18–25-year-old males. Methods: We develop a benefit–risk model, extending the FDA’s, that can stratify benefits and risks of vaccination by prior-infection and comorbidity status. We use the FDA’s framework but apply our model to account for benefits derived from prior COVID infection, while also accounting for finer age stratification in COVID-hospitalization rates, incidental hospitalizations (those of patients who test positive for COVID but receive treatment for something else), more realistic projections of Omicron-infection rates, and more accurate VAM/P rates. Results: With hospitalizations as the principal endpoint of the analysis (those prevented by vaccination vs. those caused by VAM/P), our model finds vaccine risks outweighed benefits for 18–25-year-old males, except in scenarios projecting implausibly high Omicron-infection prevalence. Our assessment suggests that mRNA-1273 vaccination of 18–25-year-old males generated between 8% and 52% more hospitalizations for VAM/P compared to COVID hospitalizations prevented (over a five-month period of vaccine protection assumed by the FDA). The preceding assessment uses model inputs based on data available at the time of the FDA’s mRNA-1273 assessment. Moreover, these inputs as well as model outputs are validated by subsequently available data. Conclusions: The outcome of a vaccine benefit–risk assessment may be dramatically impacted by accounting for the benefits derived from prior infection by the vaccine-targeted disease. To increase public confidence in vaccines and thereby reduce vaccine hesitancy, public-health agencies should employ benefit–risk models capable of supporting stratification of vaccination recommendations not only based on age and sex but also on prior-infection and comorbidity status.
We present a heavy traffic analysis for a single server queue with renewal arrivals and generally distributed i.i.d. service times, in which the server employs the Shortest Remaining Processing Time (SRPT) policy. Under typical heavy traffic assumptions, we prove a diffusion limit theorem for a measure-valued state descriptor, from which we conclude a similar theorem for the queue length process. These results allow us to make some observations on the queue length optimality of SRPT. In particular, they provide the sharpest illustration of the well-known tension between queue length optimality and quality of service for this policy.
We consider point-to-point directed paths in a random environment on the two-dimensional integer lattice. For a general independent environment under mild assumptions we show that the quenched energy of a typical path satisfies a central limit theorem as the mesh of the lattice goes to zero. Our proofs rely on concentration of measure techniques and some combinatorial bounds on families of paths.
We consider a sequence of single-server queueing models operating under a service policy that incorporates batches into processor sharing: arriving jobs build up behind a gate while waiting to begin service, while jobs in front of the gate are served according to processor sharing. When they have been completed, the waiting jobs move in front of the gate and the cycle repeats. We model this system with a pair of measure valued processes describing the jobs in front of and behind the gate. Under mild asymptotically critical conditions and a law-of-large-numbers scaling, we prove that the pair of measure-valued processes converges in distribution to an easily described limit, which has an interesting periodic dynamics.