Negi and Yilmaz: Optimal Scheduling in QA Forums and Capacity

Rohit Negi and Mustafa Yilmaz submitted a paper to arXiv on 18 June 2026 titled "Optimal Scheduling in a Question-Answering Forum of Knowledge Workers," modeling a future forum that employs topical experts and designing schedulers that reach the system's capacity. The 14-page paper (4 figures) frames the request-answer interaction as a queuing system, calculates the capacity needed to keep the system stable, and studies how collaboration among experts affects throughput.

What did the paper do?

The paper builds a queuing model for a question-answering forum with knowledge workers, computes the system capacity for stability, and designs schedulers that achieve that capacity. The authors explicitly consider requests partitioned by topic, experts with topic-dependent expertise levels, and schedulers that assign incoming requests to experts; they then derive capacity limits and present schedulers that meet those limits. The work also includes an investigation of how collaboration between experts can potentially increase capacity.

Negi and Yilmaz present their results in a formal computer-science style: the submission is listed on arXiv as arXiv:2606.19759, labeled in the cs.AI and cs.SI categories, and carries an arXiv-issued DOI via DataCite. The posted version is v1, submitted 18 Jun 2026 (file size 605 KB).

How does the scheduling model work?

The model treats the forum as a queuing system where schedulers assign topic-specific requests to experts who have different expertise levels across topics. The opening description states that requests in different topics are assigned to experts "who may be able to answer them according to their expertise levels in different topics," and the schedulers decide those assignments to control system stability.

From that setup the paper proceeds to two technical tasks: first, it calculates the system capacity required to handle the incoming requests while keeping the queues stable; second, it designs schedulers that achieve that capacity. The authors then extend the baseline model to allow collaboration between experts and quantify how that collaboration can change capacity. The abstract and metadata emphasize these three linked contributions: capacity calculation, capacity-achieving scheduler design, and analysis of collaboration's impact.

Why it matters

If QA forums evolve from volunteer contributors to systems that employ knowledge workers, capacity and scheduling become operational constraints, not abstract concerns. A scheduler that achieves capacity determines how many requests the system can handle without backlog growth. The paper's focus on topic-based assignment and expertise levels directly addresses the matching problem that limits throughput. The additional finding that expert collaboration can potentially increase capacity suggests design levers beyond simple one-to-one assignment: systems can alter who works together, not just who gets which request.

What to watch

Watch for follow-up versions and citations to see how the theoretical capacity results hold up in empirical settings, and for any code or simulations tied to the paper (the arXiv entry links to PDF and TeX source). Also monitor whether researchers extend the model to concrete platforms or to empirical measures of expertise and collaboration effectiveness.

Additional bibliographic facts: the submission is 14 pages long with 4 figures and is available as arXiv:2606.19759 [cs.AI], submitted 18 Jun 2026. The authors are Rohit Negi and Mustafa Yilmaz.