Constrained minimization of a function of many variables
Technical Report
·
OSTI ID:4305864
Powell's direct search method for finding the minimum or maximum of a function of many variables has been modified to allow inequality and equality constraints ---either or both--to be imposed. The modification involves a tnansformation of the constrained problem into an artificial unconstrained minimization problem by the addition of penalty functions to the actual function being minimized. The technique is of general applicability in optimization problems encountered in engineering and physics. A unique feature is the automatic generation of weighting factors which provide scaling of the constraint function being minimized. Automatic scaling simplifies preparation of the problem and accelerates convergence. (auth)
- Research Organization:
- California Univ., Livermore (USA). Lawrence Livermore Lab.
- DOE Contract Number:
- W-7405-ENG-48
- NSA Number:
- NSA-29-031621
- OSTI ID:
- 4305864
- Report Number(s):
- UCRL--51517
- Country of Publication:
- United States
- Language:
- English
Similar Records
Differentiable exact penalty functions via Hestenis-Powell-Rockafellar`s augmented Lagrangian function
A robust trust-region algorithm with a non-monotonic penality parameter scheme for constrained optimization
Simple procedures for imposing constraints for nonlinear least squares optimization
Conference
·
Fri Dec 30 23:00:00 EST 1994
·
OSTI ID:36219
A robust trust-region algorithm with a non-monotonic penality parameter scheme for constrained optimization
Conference
·
Fri Dec 30 23:00:00 EST 1994
·
OSTI ID:35981
Simple procedures for imposing constraints for nonlinear least squares optimization
Conference
·
Sat Dec 30 23:00:00 EST 1995
·
OSTI ID:170086