A fast and space-economical algorithm for length-limited coding
Authors:Jyrki Katajainen, Alistair Moffat, and Andrew Turpin
Published in:Proceedings of the 6th Annual International Symposium on Algorithms and Computation, Lecture Notes in Computer Science 1004, Springer-Verlag (1995), 12-21
Full text:<pdf.gif>PDF (648 kB)  
DOI:10.1007/BFb0015404
Copyright:© Springer-Verlag
Abstract: The optimal prefix code problem is to determine a vector of integer codeword lengths l = [li  |  i ∈ {1… n}], given an n-vector of symbol frequencies p = [pi  |  i ∈ {1… n}], such that ∑i=1n 2li ≤ 1, and ∑i=1n li pi is minimised. An extension of the problem is the optimal length-limited prefix code problem, in which the further constraint liL is imposed, for all i ∈ {1… n} and some integer L ≥ ⌈log2 n⌉. The package-merge algorithm of Larmore and Hirschberg generates length-limited codes in O(nL) time using O(n) words of auxiliary space. In this paper we show how the size of this work space can be reduced to O(L2). This represents a substantial improvement, since for practical purposes L is Θ(log n). The modified algorithm can also be applied to the generation of optimal prefix codes, and it reduces the memory space required for some instances of that problem.
BibLATEX:
@inproceedings{KMT1995bC,
  author = {Jyrki Katajainen and Alistair Moffat and Andrew Turpin},
  title = {A fast and space-economical algorithm for length-limited coding},
  booktitle = {Proceedings of the 6th Annual International Symposium on
    Algorithms and Computation},
  series = {Lecture Notes in Computer Science},
  volume = {1004},
  publisher = {Springer-Verlag},
  year = {1995},
  pages = {12--21},
}
This page was generated by Jyrki Katajainen <jyrki@di.ku.dk> on 31.12.2011.