Papers › Welfare-Maximizing Pooled Testing
Welfare-Maximizing Pooled Testing
Simon Finster, Michelle González Amador, Edwin Lock, Francisco Marmolejo-Cossío, Evi Micha, Ariel D. Procaccia
The archive published only this paper's code-link row. Authors, date and abstract are from arXiv's metadata (CC0), read from the Kaggle arXiv metadata snapshot of 2026-09-12 where its title matched the archive's; the title is the archive's.
Pooled testing increases the reach of scarce diagnostic resources, but optimally composing pools for individuals differing in infection risk and the utility they derive from a negative test is combinatorially challenging. We study the problem of maximizing the expected welfare of individuals cleared by a negative result, given a testing budget. Assigning a sample to several pools can raise welfare but is operationally burdensome; we show the restriction to non-overlapping allocations costs at most a factor of two for any budget or population, less under a pool-size cap at high health probabilities, and nothing when no health probability exceeds one-half. Welfare decomposes across non-overlapping pools, whereas evaluating overlapping allocations is #P-hard for pools of three or more. Finding optimal allocations is NP-hard and admits no FPTAS, with or without overlap, unless P = NP. We provide single-test routines and greedy algorithms with constant-factor guarantees. On real-world data, greedy achieves over 99% of optimal non-overlapping welfare in milliseconds, against hours for exact benchmarks. In a randomized field experiment at a Mexican research institute, our mechanism conditioned campus access on negative qPCR results. Relative to unrestricted access, we found no statistical evidence of adverse effects on participants' performance, learning, or mental health.
Code
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Results from the paper archive 2025-07-28
No leaderboard rows for this paper in the archive.
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections