若干在线排序新问题的研究.pptVIP

若干在线排序新问题研究 何 勇 浙江大学数学系 Outline Introduction Semi-online scheduling with inexact partial information Scheduling with machine cost Scheduling Problem n independent jobs m parallel identical machines The objective is to minimize the maximum machine load (makespan) Jobs and machines are available at time zero Preemption Preemption allows to split a job and spread its size over several machines. At any time, each job can be processed by at most one machine Three scheduling versions Offline Have full information on the job data before scheduling Online Schedule jobs irrevocably onto the machines as soon as they are given, without any knowledge about jobs that follow later on Semi-online Some partially additional information about jobs is available, or additional algorithmic extension is possible Algorithm measure The quality of the performance of an (semi-) online algorithm A is measured by its competitive ratio An (semi-) online problem has a lower bound if no (semi-) online algorithm has a competitive ratio smaller than An (semi-) online algorithm is called optimal if its competitive ratio matches the lower bound Lower Bound of Online algorithms The competitive ratio of LS for is , and LS is optimal for Best results for Competitive ratio 1.9201 Lower bound 1.88 The optimal algorithm for has a competitive ratio of Reference Graham, R. L.: Bounds for certain multi-processing anomalies. The Bell System Technical Journal, 45, 1563-1581, 1966. Faigle, U. , Kern, W. , Turan, G. : On the performance of online algorithms for partition problems. Acta Cybernetica, 9, 107-119, 1989. Chen, B., van Vliet, A., Woeginger, G.: An optimal algorithm for preemptive on-line scheduling, Operations Research Letters, 18, 127-131, 1995. Fleischer, R. , Wahl, M. : Online scheduling revisited. Proc. 8th Annual European Symposium on Algorithms, 202-210, 2000. Rudin III, J.F., Chandrasekaran, R.: Improved bounds for the online