首页> 外文会议>Computing and combinatorics >External Memory Soft Heap, and Hard Heap, a Meldable Priority Queue
【24h】

External Memory Soft Heap, and Hard Heap, a Meldable Priority Queue

机译:外部存储器软堆和硬堆,可混合的优先级队列

获取原文
获取原文并翻译 | 示例

摘要

An external memory version of soft heap that we call "External Memory Soft Heap" (EMSH) is presented. It supports Insert, Findmin, Deletemin and Meld operations and as in soft heap, it guarantees that the number of corrupt elements in it is never more than ∈N, where N is the total number of items inserted in it, and ∈ is a parameter of it called the error-rate. For N = O(Bm~(M/2(B+1/2m))), the amortised I/O complexity of an Insert is O(1/B log_m 1/∈), where M is the size of the main memory, B is the size of a disk block and m = M/B. Findmin, Deletemin and Meld all have non-positive amortised I/O complexities. When we choose an error rate ∈ < 1/N, EMSH stays devoid of corrupt nodes, and thus becomes a meldable priority queue that we call "hard heap". The amortised I/O complexity of an Insert, in this case, is O(1/B log_m N/B), over a sequence of operations involving N Inserts. Findmin, Deletemin and Meld all have non-positive amortised I/O complexities. If the inserted keys are all unique, a Delete (by key) operation can also be performed at an amortised I/O complexity of O(1/B log_m N/B). A balancing operation performed at appropriate intervals on a hard heap ensures that the number of I/Os performed by a sequence of S operations on it is O(S/B + 1/B ∑_(i=1)~S log_m N-i/B), where N_i is the number of elements in the heap before the ith operation.
机译:介绍了我们称为“外部内存软堆”(EMSH)的软堆的外部内存版本。它支持Insert,Findmin,Deletemin和Meld操作,并且像在软堆中一样,它保证其中损坏元素的数量永远不超过∈N,其中N是插入其中的项目总数,而∈是参数其中的一个称为错误率。对于N = O(Bm〜(M / 2(B + 1 / 2m))),插入的摊销I / O复杂度为O(1 / B log_m 1 /∈),其中M是主体的大小内存,B是磁盘块的大小,m = M / B。 Findmin,Deletemin和Meld都具有非正数摊销的I / O复杂度。当我们选择错误率∈<1 / N时,EMSH会保持无损坏节点的状态,因此成为可熔优先级队列,我们​​称之为“硬堆”。在涉及N个插入的一系列操作中,插入的摊销I / O复杂度在这种情况下为O(1 / B log_m N / B)。 Findmin,Deletemin和Meld都具有非正数摊销的I / O复杂度。如果插入的键都是唯一的,则还可以按O(1 / B log_m N / B)的摊销I / O复杂度执行Delete(按键)操作。在硬堆上以适当的时间间隔执行的平衡操作可确保通过一系列S操作对其执行的I / O数量为O(S / B + 1 / B ∑_(i = 1)〜S log_m Ni / B),其中N_i是第i个操作之前堆中元素的数量。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号