Pure Tail Constraints for Online Problems

2026-09-28

AI Summary

This study addresses the inherent tension between expected performance and worst-case competitiveness in online optimization, where tail risk control under adaptive settings has long remained limited. Moving beyond non-adaptive frameworks, this work establishes the first purely tail-risk-constrained framework for adaptive online algorithms. By integrating competitive analysis with Pareto optimality theory, it systematically investigates the trade-off mechanisms between expected and worst-case guarantees. Key contributions include precisely deriving the Pareto front for search problems to characterize optimal trade-off curves, and overcoming longstanding lower-bound challenges in adaptive problems such as TCP acknowledgment by proving significantly non-trivial bounds. These results provide a novel theoretical foundation for designing robust online decision-making systems.

Abstract

Controlling tail risk is an important objective in online optimization, and recently it has been studied in the context of competitive analysis. Continuing this line of research, we investigate pure tail constraints, which capture the tradeoff between expected and worst-case competitiveness. For two fundamental search problems, online bidding and line search, we derive the Pareto-optimal frontiers of this tradeoff. We then investigate another classic problem, TCP acknowledgment, which has structure similar to the iterated ski rental problem. There, we construct an algorithm whose tradeoff coincides with the known Pareto-optimal tradeoff for ski rental. The lower bounds for this problem are substantially more involved as the problem exhibits adaptive structure: an online algorithm observes requests of the adversary (packet arrivals) and may adaptively adjust its actions (acknowledgments) on this basis. We emphasize that all previous work on tail risk in the context of competitive analysis was restricted to non-adaptive problems, where the feedback given to an algorithm was essentially limited to a binary indicator of whether the algorithm has succeeded or not. Nonetheless, we identify a set of constraints implied by tail bounds in this adaptive setting, and show that they imply a nontrivial lower bound on the TCP acknowledgment problem.

Problem

• online optimization
• tail risk
• competitive analysis
• pure tail constraints
• adaptive problems

Innovation

• pure tail constraints
• competitive analysis
• Pareto-optimal frontiers
• adaptive structure
• TCP acknowledgment