English

Towards PTAS for Precedence Constrained Scheduling via Combinatorial Algorithms

Data Structures and Algorithms 2020-04-28 v2

Abstract

We study the classic problem of scheduling nn precedence constrained unit-size jobs on m=O(1)m = O(1) machines so as to minimize the makespan. In a recent breakthrough, Levey and Rothvoss \cite{LR16} developed a (1+ϵ)(1+\epsilon)-approximation for the problem with running time exp(exp(O(m2ϵ2log2logn)))\exp\Big(\exp\Big(O\big(\frac{m^2}{\epsilon^2}\log^2\log n\big)\Big)\Big), via the Sherali-Adams lift of the basic linear programming relaxation for the problem by exp(O(m2ϵ2log2logn))\exp\Big(O\big(\frac{m^2}{\epsilon^2}\log^2\log n\big)\Big) levels. Garg \cite{Garg18} recently improved the number of levels to logO(m2/ϵ2)n\log ^{O(m^2/\epsilon^2)}n, and thus the running time to exp(logO(m2/ϵ2)n)\exp\big(\log ^{O(m^2/\epsilon^2)}n\big), which is quasi-polynomial for constant mm and ϵ\epsilon. In this paper we present an algorithm that achieves (1+ϵ)(1+\epsilon)-approximation for the problem with running time nO(m4ϵ3log3logn)n^{O\left(\frac{m^4}{\epsilon^3}\log^3\log n\right)}, which is very close to a polynomial for constant mm and ϵ\epsilon. Unlike the algorithms of Levey-Rothvoss and Garg, which are based on linear-programming hierarchy, our algorithm is purely combinatorial. For this problem, we show that the conditioning operations on the lifted LP solution can be replaced by making guesses about the optimum schedule.

Keywords

Cite

@article{arxiv.2004.01231,
  title  = {Towards PTAS for Precedence Constrained Scheduling via Combinatorial Algorithms},
  author = {Shi Li},
  journal= {arXiv preprint arXiv:2004.01231},
  year   = {2020}
}
R2 v1 2026-06-23T14:37:21.533Z