Title
Approximation of a retrieval problem for parallel disks
Abstract
We study a number of retrieval problems that relate to effectively using the throughput of parallel disks. These problems can be formulated as assigning a maximum number of jobs to machines of capacity two, where jobs are of size one or two that must satisfy assignment restrictions. We prove that the LP-relaxation of an integer programming formulation is half-integral, and derive an interesting persistency property. In addition, we derive 2/3-approximation results for two types of retrieval problems.
Year
DOI
Venue
2003
10.1007/3-540-44849-7_23
CIAC
Keywords
Field
DocType
3-approximation result,integer programming formulation,interesting persistency property,assignment restriction,retrieval problem,parallel disk,maximum number,lp relaxation,satisfiability
Speech processing,Combinatorics,Computer science,Algorithm,Image processing,Integer programming,Throughput,Linear programming relaxation,Channel capacity
Conference
Volume
ISSN
ISBN
2653
0302-9743
3-540-40176-8
Citations 
PageRank 
References 
4
0.60
15
Authors
3
Name
Order
Citations
PageRank
Joep Aerts113417.32
Jan Korst217519.94
Frits C. R. Spieksma3244.59