Asymptotic Complexity Estimates for Probabilistic Programs and their VASS Abstractions

Michal Ajdarów,Antonín Kučera

CoRR(2023)

引用 0|浏览4
暂无评分
摘要
The standard approach to analyzing the asymptotic complexity of probabilistic programs is based on studying the asymptotic growth of certain expected values (such as the expected termination time) for increasing input size. We argue that this approach is not sufficiently robust, especially in situations when the expectations are infinite. We propose new estimates for the asymptotic analysis of probabilistic programs with non-deterministic choice that overcome this deficiency. Furthermore, we show how to efficiently compute/analyze these estimates for selected classes of programs represented as Markov decision processes over vector addition systems with states.
更多
查看译文
关键词
probabilistic programs,complexity,estimates
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要