an:00399279
Zbl 0781.65050
Berger, M. F.
Hierarchical gradient methods for nonlinear LSQ problems
EN
J. Sci. Comput. 7, No. 3, 197-228 (1992).
0885-7474 1573-7691
1992
j
65K05 90C30
global convergence; decomposition; hierarchical optimization; gradient algorithm; nonlinear least squares problems
The author proposes a variant of the gradient algorithm that consists in a succession of cycles. On an iteration of the cycle, only a subset of the variables is changed; all variables must have changed at the end of a cycle. He applies his idea to nonlinear least squares problems.
J.F.Bonnans (Le Chesnay)