New approaches to linear and nonlinear programming
During the last twelve months, research has concentrated on barrier- function methods for linear programming (LP) and quadratic programming (QP). Some ground-work for the application of barrier methods to nonlinearly constrained problems has also begun. In our previous progress report we drew attention to the difficulty of developing robust implementations of barrier methods for LP. We have continued to refine both the primal algorithm and the dual algorithm. We still do not claim that the barrier algorithms are as robust as the simplex method; however, the dual algorithm has solved all the problems in our extensive test set. We have also gained some experience with using the algorithms to solve aircrew scheduling problems.
- Research Organization:
- Stanford Univ., CA (United States). Dept. of Operations Research
- Sponsoring Organization:
- USDOE; USDOE, Washington, DC (United States)
- DOE Contract Number:
- FG03-87ER25030
- OSTI ID:
- 5254075
- Report Number(s):
- DOE/ER/25030-T1; ON: DE92014086
- Country of Publication:
- United States
- Language:
- English
Similar Records
Continuation methods for linear and quadratic programming
Issues of piecewise linearity in dual and primal-dual methods for bounded-variable linear programs