A fully-pipelined solutions constructor for dynamic programming problems

  • Jean Frédéric Myoupo
Parallel Processing And Systems
Part of the Lecture Notes in Computer Science book series (LNCS, volume 497)


The problem of designing modular linear systolic arrays for dynamic programming was raised in [9]. An attempt to solve this problem appeared in [8]. Unfortunately, the size of the local memory of each processor and the time delay between two consecutive processors of this array (in [8]) are still depending on n as in [9]. Moreover, these two arrays (in [8, 9]) are partially pipelined. Some elements have to be initially stored in the array. However, the difference between these two linear arrays is that the one in [8] is faster and more easy to handle than in [9]. In this paper we discuss a way of designing fully — pipelined modular linear systolic arrays for dynamic programming. The algorithm we obtain requires n2+1 processors (or simply cells) and 6n2-n-3 time steps for its execution: each cell has a local memory of size 1 and the time delay between two consecutive cells of the array is constant. As far as the author knows, It is the only fully-pipelined modular linear systolic array algorithm for dynamic programming appearing in the literature for the moment.

Key Words

Design of Algorithms Parallel Algorithms Linear Systolic Arrays Modular Arrays Complexity Dynamic Programming 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    K. Q. Brown: "Dynamic programming in computer science", CMU Tech. Report (February 1979).Google Scholar
  2. [2]
    A. L. Fischer and H. T. Kung: "Synchronizing large VLSI processor arrays", Proc. Tenth Annual IEEE/ACM Symposium on Computer Architecture, June 1983, pp. 54–58.Google Scholar
  3. [3]
    L. J. Guibas, H. T. Kung and C. D. Thompson:"Direct VLSI implementation of combinatorial algorithms", Proc. Conf. very Large Scale Integration: Architecture, Design, Fabrication, California Institute of Technology (January 1979), PP. 509–525.Google Scholar
  4. [4]
    D. E. Knuth: "The Art of computer Programming", Vol. 3 Sorting and Searching, Addison-Wesley (1973).Google Scholar
  5. [5]
    H. T. Kung: "Why systolic architecture?", IEEE Computer 15 (1) January, 1980, PP. 37–46.Google Scholar
  6. [6]
    F. T. Leighton and C. E. Leiserson:"Wafer-scale integration of systolic arrays", Proc. Twenty third Symp. Foundations of Computer Science, November 1982, PP. 297–311.Google Scholar
  7. [7]
    B. Louka and M. Tchuente:"Dynamic programming on two-dimensional systolic arrays", Inform. Process. Lett. 29, (1988), PP. 97–104.Google Scholar
  8. [8]
    J. F. Myoupo:" Dynamic programming on linear systolic arrays", Proc. 28 th Annu. Allerton Conf. Contr., Commun., Comput. PP. 254–261, Oct. 1990.Google Scholar
  9. [9]
    I. V. Ramakrishnan and P. J. Varman:" Dynamic programming and transitive closure on linear pipelines", Proc. Int. Conf. Parallel Processing, St. Charles, August 1984.Google Scholar
  10. [10]
    F. L. Van Scoy:"The parallel recognition of classes of graphs", IEEE-TC, Vol. C-29, No.7 (July 1980), PP. 563–570.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1991

Authors and Affiliations

  • Jean Frédéric Myoupo
    • 1
  1. 1.L.R.I, Bat. 490Université Paris-Sud et URA 410 CNRSOrsayFrance

Personalised recommendations