Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
1142446 | Operations Research Letters | 2014 | 6 Pages |
Abstract
Given nn independent integer-valued random variables X1,X2,…,XnX1,X2,…,Xn and an integer CC, we study the fundamental problem of computing the probability that the sum X=X1+X2+⋯+Xn is at most CC. We assume that each random variable XiXi is implicitly given by an oracle OiOi, which given two input integers n1,n2n1,n2 returns the probability of n1≤Xi≤n2n1≤Xi≤n2. We give the first deterministic fully polynomial-time approximation scheme (FPTAS) to estimate the probability up to a relative error of 1±ϵ1±ϵ. Our algorithm is based on the technique for approximately counting knapsack solutions, developed in Gopalan et al. (2011).
Related Topics
Physical Sciences and Engineering
Mathematics
Discrete Mathematics and Combinatorics
Authors
Jian Li, Tianlin Shi,