Reproduce “Learning to Emulate an Expert Projective Cone Scheduler” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Neal Master
The authors introduce an online method that learns to emulate a queue scheduler from observed backlogs and expert scheduling choices. It estimates a symmetric priority matrix and updates weights multiplicatively, including an algorithm for an unknown horizon. A theoretical bound shows average emulation loss tends to zero; in a two-queue simulation, decisions eventually match, without guaranteeing stability or real-world benefit.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.