Exact Sensitivity Analysis Using Augmented Lagrangians.
Interim technical rept.,
GEORGE WASHINGTON UNIV WASHINGTON D C INST FOR MANAGEMENT SCIENCE AND ENGINEERING
Pagination or Media Count:
This paper reviews several recently developed sensitivity analysis results for a general class of parametric nonlinear programming problems and develops formulas for calculating the sensitivity measures, i.e., the partial derivatives of a Kuhn-Tucker triple and the optimal value function, using an augmented Lagrangian. The theory, developed in the general case in terms of the ordinary Lagrangian and extended to allow the utilization of traditional penalty functions, is applied. Recently obtained results using a different manner of proof are shown to follow immediately using the present approach. Furthermore, the relationship of the calculations to those involving the usual Lagrangian are made explicit, hence all the results obtained previously in terms of the usual Lagrangian are immediately applicable. Author
- Numerical Mathematics