为了解决空间数据流中任意形状簇的聚类问题,提出了一种基于密度的空间数据流在线聚类算法(On-line density-based clustering algorithm for spatial data stream,OLDStream),该算法在先前聚类结果上聚类增量空间数据,仅对新增空间点...为了解决空间数据流中任意形状簇的聚类问题,提出了一种基于密度的空间数据流在线聚类算法(On-line density-based clustering algorithm for spatial data stream,OLDStream),该算法在先前聚类结果上聚类增量空间数据,仅对新增空间点及其满足核心点条件的邻域数据做局部聚类更新,降低聚类更新的时间复杂度,实现对空间数据流的在线聚类.OLDStream算法具有快速处理大规模空间数据流、实时获取全局任意形状的聚类簇结果、对数据流的输入顺序不敏感、并能发现孤立点数据等优势.在真实数据和合成数据上的综合实验验证了算法的聚类效果、高效率性和较高的可伸缩性,同时实验结果的统计分析显示仅有4%的空间点消耗最坏运行时间,对每个空间点的平均聚类时间约为0.033ms.展开更多
讨论单机、平行批、批容量无界、最小化最大完工时间的在线排序问题.对该排序问题,Zhang等人(G.Zhang,X.Cai and C.K.Wong,On-line algorithms for minimizing makespan on batch processing machines,NavalResearch Logistics,48(2001)...讨论单机、平行批、批容量无界、最小化最大完工时间的在线排序问题.对该排序问题,Zhang等人(G.Zhang,X.Cai and C.K.Wong,On-line algorithms for minimizing makespan on batch processing machines,NavalResearch Logistics,48(2001),241-258.)和Deng等人(X.Deng,C.K.Poon and Y.Z.Zhang,Approximation algo-rithms in batch processing,Journal of Combinatorial Optimization,7(2003),247-257.)两组作者分别独立地给出了同一个竞争比为(5+1)/2的在线算法,并证明该在线算法是最佳可能的.在他们的算法中,在每一批中的加工时间最大的工件,不妨设其准备时间为r而加工时间为p,将被滞后到(1+α)r+αp时刻以后加工,其中α=(5-1)/2.对同一问题设计了一个修订的在线算法,其中加工时间为p的工件只需要滞后到αp时刻.该在线算法仍然是最佳可能的,并且在一定意义下,该在线算法是渐近最优的.展开更多
In this paper we consider an on-line scheduling problem, where jobs with similar processing times within [1, r] arrive one by one to be scheduled in an on-line setting on two identical parallel processors without pree...In this paper we consider an on-line scheduling problem, where jobs with similar processing times within [1, r] arrive one by one to be scheduled in an on-line setting on two identical parallel processors without preemption. The objective is to nlinimize makespan. We devise a randomized on-line algorithm for this problem along with a lower bound.展开更多
In real world situations, most scheduling problems occur neither as complete off-line nor as complete on-line models. Most likely, a problem arises as an on-line model with some partial information. In this article, w...In real world situations, most scheduling problems occur neither as complete off-line nor as complete on-line models. Most likely, a problem arises as an on-line model with some partial information. In this article, we consider such a model. We study the scheduling problem P(n1,n2), where two groups of jobs are to be scheduled. The first job group is available beforehand. As soon as all jobs in the first group are assigned, the second job group appears. The objective is to minimize the longest job completion time (makespan). We show a lower bound of 3/2 even for very special cases. Best possible algorithms are presented for a number of cases. Furthermore, a heuristic is proposed for the general case. The main contribution of this paper is to discuss the impact of the quantity of available information in designing an on-line algorithm. It is interesting to note that the absence of even a little bit information may significantly affect the performance of an algorithm.展开更多
文摘为了解决空间数据流中任意形状簇的聚类问题,提出了一种基于密度的空间数据流在线聚类算法(On-line density-based clustering algorithm for spatial data stream,OLDStream),该算法在先前聚类结果上聚类增量空间数据,仅对新增空间点及其满足核心点条件的邻域数据做局部聚类更新,降低聚类更新的时间复杂度,实现对空间数据流的在线聚类.OLDStream算法具有快速处理大规模空间数据流、实时获取全局任意形状的聚类簇结果、对数据流的输入顺序不敏感、并能发现孤立点数据等优势.在真实数据和合成数据上的综合实验验证了算法的聚类效果、高效率性和较高的可伸缩性,同时实验结果的统计分析显示仅有4%的空间点消耗最坏运行时间,对每个空间点的平均聚类时间约为0.033ms.
文摘讨论单机、平行批、批容量无界、最小化最大完工时间的在线排序问题.对该排序问题,Zhang等人(G.Zhang,X.Cai and C.K.Wong,On-line algorithms for minimizing makespan on batch processing machines,NavalResearch Logistics,48(2001),241-258.)和Deng等人(X.Deng,C.K.Poon and Y.Z.Zhang,Approximation algo-rithms in batch processing,Journal of Combinatorial Optimization,7(2003),247-257.)两组作者分别独立地给出了同一个竞争比为(5+1)/2的在线算法,并证明该在线算法是最佳可能的.在他们的算法中,在每一批中的加工时间最大的工件,不妨设其准备时间为r而加工时间为p,将被滞后到(1+α)r+αp时刻以后加工,其中α=(5-1)/2.对同一问题设计了一个修订的在线算法,其中加工时间为p的工件只需要滞后到αp时刻.该在线算法仍然是最佳可能的,并且在一定意义下,该在线算法是渐近最优的.
文摘In this paper we consider an on-line scheduling problem, where jobs with similar processing times within [1, r] arrive one by one to be scheduled in an on-line setting on two identical parallel processors without preemption. The objective is to nlinimize makespan. We devise a randomized on-line algorithm for this problem along with a lower bound.
基金Research partially supported by a Hong Kong Government RGC Earmarked Grant.Ref.No.CUHK356/96E Research partially supported by National 973 Fundamental Research Project of china and National Natural Science Foundation of China (19801032)
文摘In real world situations, most scheduling problems occur neither as complete off-line nor as complete on-line models. Most likely, a problem arises as an on-line model with some partial information. In this article, we consider such a model. We study the scheduling problem P(n1,n2), where two groups of jobs are to be scheduled. The first job group is available beforehand. As soon as all jobs in the first group are assigned, the second job group appears. The objective is to minimize the longest job completion time (makespan). We show a lower bound of 3/2 even for very special cases. Best possible algorithms are presented for a number of cases. Furthermore, a heuristic is proposed for the general case. The main contribution of this paper is to discuss the impact of the quantity of available information in designing an on-line algorithm. It is interesting to note that the absence of even a little bit information may significantly affect the performance of an algorithm.
