Probabilistic Analysis of a Heuristics for the Dual Bin Packing Problem
暂无分享,去创建一个
Abstract In this paper we study the probabilistic behavior of the first fit increasing heuristics for the dual bin packing problem. In this problem we seek to maximize the number of items that can be packed into m unit capacity bins. We show that when the items to be packed are independent and identically distributed, then the relative error of first fit increasing tends to zero in probability as the number of items tends to infinity. We also consider an on-line variant of this problem and propose a simple heuristics whose relative error tends to zero in probability as the number of items tends to infinity under the assumption that the item sizes are independently and identically distributed according to a distribution with finite mean and with zero in its compact support.
[1] R. Weber,et al. Optimal selection of stochastic intervals under a sum constraint , 1987 .
[2] Joseph Y.-T. Leung,et al. Combinatorial analysis of an efficient algorithm for processor and storage allocation , 1977, 18th Annual Symposium on Foundations of Computer Science (sfcs 1977).
[3] H. Chernoff. A Measure of Asymptotic Efficiency for Tests of a Hypothesis Based on the sum of Observations , 1952 .
[4] Joseph Yuk-Tong Leung. Fast algorithms for packing problems. , 1977 .