Scheduling When You Do Not Know the Number of Machines
Scheduling When You Do Not Know the Number of Machines
复制标题
不知道机器数量时的调度
DOI：
10.1145/3340320
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Mingxian Zhong
中科院分区：
文献类型：
--
作者：
C. Stein;Mingxian Zhong
在计划问题中，对机器的不确定性问题存在不确定性。已知机器的数量，然后需要在计算机上安排这些设置，而无需分开以评估算法，我们介绍了α-射击算法的想法，保证在M机器上最佳时间表的任何数字M机器上返回时间表，其中最佳限制不受该集合无法分离的限制。 A（5 \ 3+ε） - 在平行机上安排的算法，以最小化makepan，并显示出无限量的特殊情况下的下限4 \ 3。渐近下限1.207。
Often in a scheduling problem, there is uncertainty about the jobs to be processed. The issue of uncertainty regarding the machines has been much less studied. In this article, we study a scheduling environment in which jobs first need to be grouped into some sets before the number of machines is known, and then the sets need to be scheduled on machines without being separated. To evaluate algorithms in such an environment, we introduce the idea of an α-robust algorithm, one that is guaranteed to return a schedule on any number m of machines that is within an α factor of the optimal schedule on m machine, where the optimum is not subject to the restriction that the sets cannot be separated. Under such environment, we give a (5\3+ε)-robust algorithm for scheduling on parallel machines to minimize makespan and show a lower bound 4\3. For the special case when the jobs are infinitesimal, we give a 1.233-robust algorithm with an asymptotic lower bound of 1.207. We also study a case of fair allocation, where the objective is to minimize the difference between the maximum and minimum machine load.
DOI：
10.1137/110844210
发表时间：
2012-05
期刊：
SIAM J. Comput.
影响因子：
--
作者：
Julián Mestre;Nicole Megow
通讯作者：
Julián Mestre;Nicole Megow
DOI：
10.4230/lipics.approx-random.2015.175
发表时间：
--
期刊：
影响因子：
--
作者：
L. Chen;N. Megow;R. Rischke;L. Stougie.
通讯作者：
L. Stougie.
DOI：
10.1007/978-3-319-24971-1_13
发表时间：
--
期刊：
影响因子：
--
作者：
N. Megow.
通讯作者：
N. Megow.