Journal of Optimization Theory and Applications

, Volume 54, Issue 3, pp 471–477

Local convergence analysis of a grouped variable version of coordinate descent

Authors

  • J. C. Bezdek
    • Department of Computer ScienceUniversity of South Carolina
  • R. J. Hathaway
    • Department of StatisticsUniversity of South Carolina
  • R. E. Howard
    • Department of MathematicsUniversity of South Carolina
  • C. A. Wilson
    • Department of MathematicsWinthrop College
  • M. P. Windham
    • Department of MathematicsUtah State University
Contributed Papers

DOI: 10.1007/BF00940196

Cite this article as:
Bezdek, J.C., Hathaway, R.J., Howard, R.E. et al. J Optim Theory Appl (1987) 54: 471. doi:10.1007/BF00940196

Abstract

LetF(x,y) be a function of the vector variablesxRn andyRm. One possible scheme for minimizingF(x,y) is to successively alternate minimizations in one vector variable while holding the other fixed. Local convergence analysis is done for this vector (grouped variable) version of coordinate descent, and assuming certain regularity conditions, it is shown that such an approach is locally convergent to a minimizer and that the rate of convergence in each vector variable is linear. Examples where the algorithm is useful in clustering and mixture density decomposition are given, and global convergence properties are briefly discussed.

Key Words

Coordinate descentlocal linear convergence
Download to read the full article text

Copyright information

© Plenum Publishing Corporation 1987