Arbitrary signal distributions in stochastic-signal ski rental

Handle arbitrary signal distributions in the stochastic-signal ski rental problem, extending the analysis beyond the uniform signal distribution whose corresponding methods are tied to that underlying distribution.

Background

The paper studies ski rental when the algorithm receives a signal whose arrival time is sampled from a known distribution scaled to the adversarial horizon. For the uniform distribution, it derives tight competitive ratios for deterministic and randomized algorithms and shows stability under small deviations from uniformity.

The authors identify extending the analysis to arbitrary signal distributions as the main open direction. They note that the achievable competitive-ratio improvement depends strongly on the distribution: distributions concentrated near small signal times can be highly informative, whereas distributions concentrated near the horizon provide little benefit. The existing methods are specifically tailored to the uniform case.

References

The main open direction is then to handle arbitrary signal distributions.

— Ski Rental with Uniform Stochastic Signals  (2610.26214 - Cardas et al., 19 Oct 2026) in Section Conclusion